# Towards Data Science

Publish AI, ML & data-science insights to a global community of data professionals.

This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles.

## LightGBM vs. a Four-Week Average for Forecasting Bike-Share Rentals

DevFeed: [LightGBM vs. a Four-Week Average for Forecasting Bike-Share Rentals](<https://devfeed.tech/articles/lightgbm-vs-a-four-week-average-for-forecasting-bike-share-rentals-146443.md>)

Original publisher: [Read original article](<https://towardsdatascience.com/lightgbm-vs-a-four-week-average-for-forecasting-bike-share-rentals/>)

Author: Nahid Ahmadvand

Published: 2026-11-20T16:30:00Z

Content type: article

Language: en

Sources: [Towards Data Science](<https://devfeed.tech/sources/towards-data-science.md>)

Topics: [Time Series](<https://devfeed.tech/topics/time-series.md>), [Feature Engineering](<https://devfeed.tech/topics/feature-engineering.md>), [Data Quality](<https://devfeed.tech/topics/data-quality.md>), [scikit-learn](<https://devfeed.tech/topics/scikit-learn.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [bike-share](<https://devfeed.tech/tags/bike-share.md>), [editor-s-picks](<https://devfeed.tech/tags/editor-s-picks.md>), [forecasting](<https://devfeed.tech/tags/forecasting.md>), [lightgbm](<https://devfeed.tech/tags/lightgbm.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [python](<https://devfeed.tech/tags/python.md>), [time-series-forecasting](<https://devfeed.tech/tags/time-series-forecasting.md>)

### AI overview

In a 48-hour-ahead Capital Bikeshare forecasting experiment, a four-week seasonal average proved difficult to beat during development: LightGBM's MAE was 49.025 versus 49.856, and paired tests did not establish a difference. On 30 later November-December forecast paths, the selected LightGBM pipeline with lagged weather had 19.6% lower MAE than the baseline, though the author did not test that difference for statistical significance and cannot identify what caused the change. The workflow reconstructed 159 missing development hours as zeros after hourly sums matched daily totals, while distinguishing recorded rentals from actual demand; it excluded nonexistent daylight-saving hours from scoring and constrained features to information available at midnight. Future observed weather sharply reduced error but was unavailable at forecast time. LightGBM relied heavily on weekly summaries, and it struggled during Hurricane Sandy. Empirical prediction intervals came close to target coverage during development but fell short in the later period. The experiment did not compare foundation models.

### Source excerpt

A 48-hour Capital Bikeshare experiment pits XGBoost, Ridge, SARIMA, and Prophet against simple baselines, with care for missing hours and leakage. The post LightGBM vs. a Four-Week Average for Forecasting Bike-Share Rentals appeared first on Towards Data Science.

## Clairic Hit 98% Accuracy on 72.8% of Questions via Confidence Scores

DevFeed: [Clairic Hit 98% Accuracy on 72.8% of Questions via Confidence Scores](<https://devfeed.tech/articles/clairic-hit-98-accuracy-on-72-8-of-questions-via-confidence-scores-146441.md>)

Original publisher: [Read original article](<https://towardsdatascience.com/clairic-hit-98-accuracy-on-728-of-questions-via-confidence-scores/>)

Author: Mariya Mansurova

Published: 2026-11-20T15:00:00Z

Content type: tutorial

Language: en

Sources: [Towards Data Science](<https://devfeed.tech/sources/towards-data-science.md>)

Topics: [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Large Language Model](<https://devfeed.tech/topics/llm.md>), [Data Science](<https://devfeed.tech/topics/data-science.md>), [ai-experiments](<https://devfeed.tech/topics/ai-experiments.md>), [Benchmark](<https://devfeed.tech/topics/benchmark.md>)

Tags: [analysts](<https://devfeed.tech/tags/analysts.md>), [arithmetic](<https://devfeed.tech/tags/arithmetic.md>), [artificial-intelligence](<https://devfeed.tech/tags/artificial-intelligence.md>), [autocomplete](<https://devfeed.tech/tags/autocomplete.md>), [based](<https://devfeed.tech/tags/based.md>), [bike](<https://devfeed.tech/tags/bike.md>), [charts](<https://devfeed.tech/tags/charts.md>), [chatgpt](<https://devfeed.tech/tags/chatgpt.md>), [classification](<https://devfeed.tech/tags/classification.md>), [cloudflare](<https://devfeed.tech/tags/cloudflare.md>), [complex](<https://devfeed.tech/tags/complex.md>), [confidence](<https://devfeed.tech/tags/confidence.md>), [confidence-interval](<https://devfeed.tech/tags/confidence-interval.md>), [control](<https://devfeed.tech/tags/control.md>), [data-science](<https://devfeed.tech/tags/data-science.md>), [decision-models](<https://devfeed.tech/tags/decision-models.md>), [easy](<https://devfeed.tech/tags/easy.md>), [editor-s-picks](<https://devfeed.tech/tags/editor-s-picks.md>), [generative-ai](<https://devfeed.tech/tags/generative-ai.md>), [jev](<https://devfeed.tech/tags/jev.md>), [llm](<https://devfeed.tech/tags/llm.md>), [llm-applications](<https://devfeed.tech/tags/llm-applications.md>), [luna](<https://devfeed.tech/tags/luna.md>), [model](<https://devfeed.tech/tags/model.md>), [science](<https://devfeed.tech/tags/science.md>), [system-one](<https://devfeed.tech/tags/system-one.md>)

### AI overview

Clairic is a fast, low-cost extraction model that accepts text and images and can return typed values with confidence scores. In tests against Luna, Clairic was faster but less accurate overall, especially on calculations; examples also included confidently incorrect extractions. It handled some chart-based extraction and returned null for missing revenue. On 1,360 questions across four languages, a 0.9 confidence cutoff yielded 98% accuracy while Clairic answered 72.8% of questions. The author suggests using it for high-confidence extraction and routing the rest to a larger model, with validation because confidence is imperfect. The tests did not establish a reason to prefer it for forecasting over traditional methods.

### Source excerpt

With a 0.9 confidence cutoff, Cloudflare's System One model handles the easy cases and routes the rest to a bigger, slower LLM like Luna. The post Clairic Hit 98% Accuracy on 72.8% of Questions via Confidence Scores appeared first on Towards Data Science.

## Five Guardrails That Limit the Damage an AI Agent Can Do

DevFeed: [Five Guardrails That Limit the Damage an AI Agent Can Do](<https://devfeed.tech/articles/five-guardrails-that-limit-the-damage-an-ai-agent-can-do-145977.md>)

Original publisher: [Read original article](<https://towardsdatascience.com/five-guardrails-that-limit-the-damage-an-ai-agent-can-do/>)

Author: Thomas Reid

Published: 2026-11-20T12:00:00Z

Content type: tutorial

Language: en

Sources: [Towards Data Science](<https://devfeed.tech/sources/towards-data-science.md>)

Topics: [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [Security](<https://devfeed.tech/topics/security.md>), [function calling](<https://devfeed.tech/topics/function-calling.md>), [Test-driven development](<https://devfeed.tech/topics/tdd.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [arguments](<https://devfeed.tech/tags/arguments.md>), [guardrail](<https://devfeed.tech/tags/guardrail.md>), [guardrails](<https://devfeed.tech/tags/guardrails.md>), [llmops](<https://devfeed.tech/tags/llmops.md>), [pydantic](<https://devfeed.tech/tags/pydantic.md>), [pytest](<https://devfeed.tech/tags/pytest.md>), [python](<https://devfeed.tech/tags/python.md>), [sqlite](<https://devfeed.tech/tags/sqlite.md>), [tests](<https://devfeed.tech/tags/tests.md>)

### AI overview

The article's central safeguard is to let Python enforce what an AI agent can do, since prompts alone cannot reliably prevent harmful actions. It treats user questions, model tool calls, and returned data as untrusted, including indirect prompt injection in database notes. The example uses Pydantic models to reject unknown fields and constrain arguments, an allowlisted dispatcher for read tools, read-only SQLite connections with limited query results, and a cap on the agent's turns. A return write stays outside the autonomous tool loop and is saved only after user confirmation; a UUID idempotency key and database uniqueness constraint prevent duplicate requests on retry. Pytest checks the deterministic controls. The example agent ignored Mina's malicious note, but the article cautions that this does not fully mitigate prompt injection; application-enforced limits reduce the damage a bad model decision can cause.

### Source excerpt

AI Agents for Beginners, Part three: strict Pydantic arguments, a tool allowlist, read-only SQLite, a step cap, and confirmed writes, all verified by five pytest tests in 0.68s. The post Five Guardrails That Limit the Damage an AI Agent Can Do appeared first on Towards Data Science.

## How to Run Claude Code and Other Coding Agents in the Cloud 24/7

DevFeed: [How to Run Claude Code and Other Coding Agents in the Cloud 24/7](<https://devfeed.tech/articles/how-to-run-claude-code-and-other-coding-agents-in-the-cloud-24-7-144203.md>)

Original publisher: [Read original article](<https://towardsdatascience.com/how-to-run-claude-code-and-other-coding-agents-in-the-cloud-247/>)

Author: Eivind Kjosbakken

Published: 2026-11-19T15:00:00Z

Content type: article

Language: en

Sources: [Towards Data Science](<https://devfeed.tech/sources/towards-data-science.md>)

Topics: [agentic-coding](<https://devfeed.tech/topics/agentic-coding.md>), [AI-assisted coding](<https://devfeed.tech/topics/ai-assisted-coding.md>), [Claude Code](<https://devfeed.tech/topics/claude-code.md>), [AI Infrastructure](<https://devfeed.tech/topics/ai-infrastructure.md>)

Tags: [agents](<https://devfeed.tech/tags/agents.md>), [ai-coding-assistants](<https://devfeed.tech/tags/ai-coding-assistants.md>), [ai-credits](<https://devfeed.tech/tags/ai-credits.md>), [automation](<https://devfeed.tech/tags/automation.md>), [aws](<https://devfeed.tech/tags/aws.md>), [claude](<https://devfeed.tech/tags/claude.md>), [claude-code](<https://devfeed.tech/tags/claude-code.md>), [cloud](<https://devfeed.tech/tags/cloud.md>), [cloud-coding](<https://devfeed.tech/tags/cloud-coding.md>), [code](<https://devfeed.tech/tags/code.md>), [coding](<https://devfeed.tech/tags/coding.md>), [coding-agent](<https://devfeed.tech/tags/coding-agent.md>), [coding-agents](<https://devfeed.tech/tags/coding-agents.md>), [computer](<https://devfeed.tech/tags/computer.md>), [control](<https://devfeed.tech/tags/control.md>), [cron](<https://devfeed.tech/tags/cron.md>), [data-science](<https://devfeed.tech/tags/data-science.md>), [full](<https://devfeed.tech/tags/full.md>), [how-to](<https://devfeed.tech/tags/how-to.md>), [llm-agents](<https://devfeed.tech/tags/llm-agents.md>), [llm-applications](<https://devfeed.tech/tags/llm-applications.md>), [run](<https://devfeed.tech/tags/run.md>), [server](<https://devfeed.tech/tags/server.md>)

### AI overview

Cloud coding sessions can keep long-running agents active when a local computer is closed or offline, reducing interruptions and the need to restart sessions. The quick option described is to set Claude Code in the Claude desktop app to use a cloud environment by default; it also reduces local RAM use, but the app limits automation such as cron jobs. For more flexibility and support for other agents and scheduled jobs, the author rents and configures a server, then connects through Paradev (formerly Superset) to work remotely. This route requires setup and occasional troubleshooting; the author estimates 4-5 hours initially and about two hours resolving issues. The article cites cheaper providers such as Hetzner and compares renting a 128 GB server at roughly $500/month with a $7,000-$8,000 Mac, as estimates from November 2026.

### Source excerpt

Two setups: flip Claude Code's cloud environment on for the quick path, or rent a server for full control. The post How to Run Claude Code and Other Coding Agents in the Cloud 24/7 appeared first on Towards Data Science.

## How to Oversee AI Agents at Scale With a Three-Tier Architecture

DevFeed: [How to Oversee AI Agents at Scale With a Three-Tier Architecture](<https://devfeed.tech/articles/how-to-oversee-ai-agents-at-scale-with-a-three-tier-architecture-143637.md>)

Original publisher: [Read original article](<https://towardsdatascience.com/how-to-oversee-ai-agents-at-scale-with-a-three-tier-architecture/>)

Author: Barr Moses

Published: 2026-11-19T13:30:00Z

Content type: opinion

Language: en

Sources: [Towards Data Science](<https://devfeed.tech/sources/towards-data-science.md>)

Topics: [agentic workflows](<https://devfeed.tech/topics/agentic-workflows.md>), [AI Architecture](<https://devfeed.tech/topics/ai-architecture.md>), [software-architecture](<https://devfeed.tech/topics/software-architecture.md>), [human review](<https://devfeed.tech/topics/human-review.md>), [LLM security](<https://devfeed.tech/topics/llm-security.md>), [scalable-architecture](<https://devfeed.tech/topics/scalable-architecture.md>)

Tags: [agent-architecture](<https://devfeed.tech/tags/agent-architecture.md>), [agent-observability](<https://devfeed.tech/tags/agent-observability.md>), [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [ai-evaluation](<https://devfeed.tech/tags/ai-evaluation.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [autonomous](<https://devfeed.tech/tags/autonomous.md>), [editor-s-picks](<https://devfeed.tech/tags/editor-s-picks.md>), [human-review](<https://devfeed.tech/tags/human-review.md>), [humans](<https://devfeed.tech/tags/humans.md>), [impact](<https://devfeed.tech/tags/impact.md>), [oversight](<https://devfeed.tech/tags/oversight.md>), [runtime](<https://devfeed.tech/tags/runtime.md>), [scale](<https://devfeed.tech/tags/scale.md>), [supervisor](<https://devfeed.tech/tags/supervisor.md>)

### AI overview

The article proposes a three-tier setup for scaling agent oversight: task-performing agents, an independent automated supervisor for runtime monitoring and reversible actions, and humans who set policy and approve high-impact changes. It argues that manual review cannot keep pace with agents and may miss silent failures, so the supervisor should use separate compute and credentials, independent telemetry, and append-only audit logs. Humans define task-level autonomy rules; the supervisor checks them against independently computed metrics, can pause or buffer a batch, and escalates irreversible fixes for approval. In the currency-conversion example, a spike in transactions using a fallback rate triggers a batch hold and a proposed schema-mapping patch for an engineer to approve. The author recommends keeping the supervisor simple, limiting it to easy-to-undo actions, and having humans spot-check its decisions.

### Source excerpt

Separate runtime oversight from human review: an autonomous supervisor handles reversible fixes, while humans approve high-impact changes. The post How to Oversee AI Agents at Scale With a Three-Tier Architecture appeared first on Towards Data Science.

## Neurosymbolic Agents Explained, With a Z3 Example in Python

DevFeed: [Neurosymbolic Agents Explained, With a Z3 Example in Python](<https://devfeed.tech/articles/neurosymbolic-agents-explained-with-a-z3-example-in-python-140794.md>)

Original publisher: [Read original article](<https://towardsdatascience.com/neurosymbolic-agents-explained-with-a-z3-example-in-python/>)

Author: Tomoharu Hayakawa

Published: 2026-11-18T15:00:00Z

Content type: article

Language: en

Sources: [Towards Data Science](<https://devfeed.tech/sources/towards-data-science.md>)

Topics: [llm-reasoning](<https://devfeed.tech/topics/llm-reasoning.md>), [agentic workflows](<https://devfeed.tech/topics/agentic-workflows.md>), [Python](<https://devfeed.tech/topics/python.md>)

Tags: [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [editor-s-picks](<https://devfeed.tech/tags/editor-s-picks.md>), [large-language-models](<https://devfeed.tech/tags/large-language-models.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [neurosymbolic-ai](<https://devfeed.tech/tags/neurosymbolic-ai.md>), [python](<https://devfeed.tech/tags/python.md>), [z3](<https://devfeed.tech/tags/z3.md>)

### AI overview

Neurosymbolic agents can verify that an answer satisfies a formal specification, but cannot guarantee that the specification captures the user's intent. The article compares four uses: Logic-LM sends solver errors back to revise a translation and outperformed prompting on five reasoning datasets; Agent-C uses an SMT solver to block policy-breaking tool calls, reporting full policy conformance on retail and airline benchmarks; Amazon's Bedrock Guardrails checks responses against multiple policy translations that must agree; and theorem-prover training uses symbolic checks as rewards. A Z3 scheduling example shows how named constraints can identify conflicting English sentences, while multiple interpretations can reveal ambiguity and prompt a user question or a more conservative schedule. The central limitation is the formalization gap: studies cited find that valid, solvable translations can still describe the wrong problem. Suggested practices include typed specifications instead of model-written solver code, traceable constraint names, multiple translations, and human review.

### Source excerpt

Starting from Air Canada's chatbot case, a tour of Logic-LM, Agent-C, and Amazon's checks shows the solver is sound but the English-to-logic step is not. The post Neurosymbolic Agents Explained, With a Z3 Example in Python appeared first on Towards Data Science.

## When Qini Rewards the Wrong Uplift Model, Use Doubly Robust Scores

DevFeed: [When Qini Rewards the Wrong Uplift Model, Use Doubly Robust Scores](<https://devfeed.tech/articles/when-qini-rewards-the-wrong-uplift-model-use-doubly-robust-scores-140795.md>)

Original publisher: [Read original article](<https://towardsdatascience.com/when-qini-rewards-the-wrong-uplift-model-use-doubly-robust-scores/>)

Author: Ivan Gorban

Published: 2026-11-18T13:30:00Z

Content type: article

Language: en

Sources: [Towards Data Science](<https://devfeed.tech/sources/towards-data-science.md>)

Topics: [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [AI Engineering](<https://devfeed.tech/topics/ai-engineering.md>), [Inference](<https://devfeed.tech/topics/inference.md>)

Tags: [ate](<https://devfeed.tech/tags/ate.md>), [bias](<https://devfeed.tech/tags/bias.md>), [bootstrap](<https://devfeed.tech/tags/bootstrap.md>), [causal-inference](<https://devfeed.tech/tags/causal-inference.md>), [confidence](<https://devfeed.tech/tags/confidence.md>), [data-science](<https://devfeed.tech/tags/data-science.md>), [dr](<https://devfeed.tech/tags/dr.md>), [editor-s-picks](<https://devfeed.tech/tags/editor-s-picks.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [marketing-analytics](<https://devfeed.tech/tags/marketing-analytics.md>), [model](<https://devfeed.tech/tags/model.md>), [ranking](<https://devfeed.tech/tags/ranking.md>), [robust](<https://devfeed.tech/tags/robust.md>), [uplift](<https://devfeed.tech/tags/uplift.md>), [uplift-modeling](<https://devfeed.tech/tags/uplift-modeling.md>)

### AI overview

In a simulation of one million customers with selective treatment assignment, treatment/control differences made calibrated Qini favor a Linear DR ranking over the known true-CATE ranking (0.0856 vs. 0.0826), even though the true ranking had greater causal gain. The authors explain that Qini uses observed outcomes and does not remove selection bias; calibration aligns scores at one reference but preserves each metric's model ordering. IPW and doubly robust (DR) gain scores were closer to known gains in the simulation. IPW requires a correct propensity model; DR requires a correct propensity model or both outcome models. Both require consistency, no unmeasured confounding, and overlap, and remain subject to estimation and finite-sample error. In this specific comparison, the authors favor a boosted S-learner as a simple starting point: its DR-gain advantage over DML was small and uncertain, while only its comparison with T showed a paired interval excluding zero. They caution that the bootstrap held fitted models fixed and that the findings come from simulations, not a guarantee for other datasets.

### Source excerpt

On 1 million simulated customers, calibrated Qini scored a Linear DR learner 0.0856 against 0.0826 for the true-CATE ranking that should have won. The post When Qini Rewards the Wrong Uplift Model, Use Doubly Robust Scores appeared first on Towards Data Science.

## How to Estimate Dose-Response Curves When DoubleML Falls Short

DevFeed: [How to Estimate Dose-Response Curves When DoubleML Falls Short](<https://devfeed.tech/articles/how-to-estimate-dose-response-curves-when-doubleml-falls-short-140793.md>)

Original publisher: [Read original article](<https://towardsdatascience.com/how-to-estimate-dose-response-curves-when-doubleml-falls-short/>)

Author: Ryan OSullivan

Published: 2026-11-18T12:00:00Z

Content type: article

Language: en

Sources: [Towards Data Science](<https://devfeed.tech/sources/towards-data-science.md>)

Topics: [doubleml](<https://devfeed.tech/topics/doubleml.md>), [Algorithms](<https://devfeed.tech/topics/algorithms.md>)

Tags: [2-2](<https://devfeed.tech/tags/2-2.md>), [2014](<https://devfeed.tech/tags/2014.md>), [2017](<https://devfeed.tech/tags/2017.md>), [2018](<https://devfeed.tech/tags/2018.md>), [2020](<https://devfeed.tech/tags/2020.md>), [2021](<https://devfeed.tech/tags/2021.md>), [2022](<https://devfeed.tech/tags/2022.md>), [2025](<https://devfeed.tech/tags/2025.md>), [aggregate](<https://devfeed.tech/tags/aggregate.md>), [aipw](<https://devfeed.tech/tags/aipw.md>), [analysis](<https://devfeed.tech/tags/analysis.md>), [app](<https://devfeed.tech/tags/app.md>), [causal-inference](<https://devfeed.tech/tags/causal-inference.md>), [data](<https://devfeed.tech/tags/data.md>), [data-science](<https://devfeed.tech/tags/data-science.md>), [doubleml](<https://devfeed.tech/tags/doubleml.md>), [editor-s-picks](<https://devfeed.tech/tags/editor-s-picks.md>), [estimator](<https://devfeed.tech/tags/estimator.md>), [hidden](<https://devfeed.tech/tags/hidden.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [response](<https://devfeed.tech/tags/response.md>), [robust](<https://devfeed.tech/tags/robust.md>), [simulations](<https://devfeed.tech/tags/simulations.md>), [statistics](<https://devfeed.tech/tags/statistics.md>), [test](<https://devfeed.tech/tags/test.md>)

### AI overview

For continuous treatments, DoubleML's usual partially linear model estimates one linear slope; it can miss nonlinear effects and becomes noisy when fitted separately to small groups. The article presents a doubly robust dose-response curve: cross-fit outcome and dose models, combine them into a pseudo-outcome, then smooth it with partial pooling across nested groups. In simulations, this method captured inverted-U and saturating effects, corrected much of the error when one nuisance model was misspecified, and improved estimates for small groups. Learner choice mattered: density weights could fail with boosting alone, while balancing weights performed better in that setting; a stack combining boosting and splines worked well on smooth data, though no learner won everywhere. For hidden confounding, difference-in-differences helped when the unmeasured trait was stable, subject to a stronger parallel-trends assumption. The author cautions that simulated confidence bands often undercovered and identifies poor dose overlap, group-level treatment, zero-dose spikes, and crossed groups as limitations.

### Source excerpt

Simulations test DoubleML against Kennedy's doubly robust estimator, with partial pooling across 200 stores and a DiD fix for hidden confounders. The post How to Estimate Dose-Response Curves When DoubleML Falls Short appeared first on Towards Data Science.

## How to Improve Coding Agent Efficiency by 30% with an Orchestrator

DevFeed: [How to Improve Coding Agent Efficiency by 30% with an Orchestrator](<https://devfeed.tech/articles/how-to-improve-coding-agent-efficiency-by-30-with-an-orchestrator-138252.md>)

Original publisher: [Read original article](<https://towardsdatascience.com/how-to-improve-coding-agent-efficiency-by-30-with-an-orchestrator/>)

Author: Eivind Kjosbakken

Published: 2026-11-17T16:30:00Z

Content type: article

Language: en

Sources: [Towards Data Science](<https://devfeed.tech/sources/towards-data-science.md>)

Topics: [Multi Agent Systems](<https://devfeed.tech/topics/multi-agent-systems.md>), [agent to agent communication](<https://devfeed.tech/topics/agent-to-agent-communication.md>), [codex](<https://devfeed.tech/topics/codex.md>), [human computer interaction](<https://devfeed.tech/topics/human-computer-interaction.md>), [UI/UX Design](<https://devfeed.tech/topics/ui-ux-design.md>)

Tags: [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [ai-coding-assistants](<https://devfeed.tech/tags/ai-coding-assistants.md>), [claude-code](<https://devfeed.tech/tags/claude-code.md>), [coding](<https://devfeed.tech/tags/coding.md>), [coding-agents](<https://devfeed.tech/tags/coding-agents.md>), [efficiency](<https://devfeed.tech/tags/efficiency.md>), [how-to](<https://devfeed.tech/tags/how-to.md>), [llm-agents](<https://devfeed.tech/tags/llm-agents.md>), [orchestrator](<https://devfeed.tech/tags/orchestrator.md>), [priority](<https://devfeed.tech/tags/priority.md>), [programming](<https://devfeed.tech/tags/programming.md>), [sessions](<https://devfeed.tech/tags/sessions.md>), [webhooks](<https://devfeed.tech/tags/webhooks.md>)

### AI overview

The author reports efficiency gains of more than 30% after routing coding-agent sessions through a centralized "herder" that surfaces and prioritizes requests for user input. The estimate is based on the author's work before and after adopting the setup. The author first used minute-by-minute polling with a cheap model such as Anthropic Haiku 5.5, then switched to webhooks that notify the herder when a session finishes or needs input; webhooks require more setup and were more efficient in this experience. A separate orchestrator agent monitors stalled sessions, attempts to resume them, and alerts the herder if it cannot. The author also uses the herder to start sessions and present questions with HTML reports. The claimed gain and reported indicator inaccuracy reflect the author's experience.

### Source excerpt

Stop hunting through sessions for pending questions: a herder agent surfaces them by priority, kept current by webhooks or cheap Haiku polling. The post How to Improve Coding Agent Efficiency by 30% with an Orchestrator appeared first on Towards Data Science.

## Google ADK Agents That Pause, Resume, and Wait Hours for a Human

DevFeed: [Google ADK Agents That Pause, Resume, and Wait Hours for a Human](<https://devfeed.tech/articles/google-adk-agents-that-pause-resume-and-wait-hours-for-a-human-137192.md>)

Original publisher: [Read original article](<https://towardsdatascience.com/google-adk-agents-that-pause-resume-and-wait-hours-for-a-human/>)

Author: Benjamin Etienne

Published: 2026-11-17T13:30:00Z

Content type: article

Language: en

Sources: [Towards Data Science](<https://devfeed.tech/sources/towards-data-science.md>)

Topics: [google-adk](<https://devfeed.tech/topics/google-adk.md>), [async](<https://devfeed.tech/topics/async.md>), [human in the loop agent](<https://devfeed.tech/topics/human-in-the-loop-agent.md>), [long-running](<https://devfeed.tech/topics/long-running.md>), [agentic workflows](<https://devfeed.tech/topics/agentic-workflows.md>), [Google Cloud Platform (GCP)](<https://devfeed.tech/topics/google-cloud.md>)

Tags: [2](<https://devfeed.tech/tags/2.md>), [3](<https://devfeed.tech/tags/3.md>), [adk](<https://devfeed.tech/tags/adk.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [asynchronous-programming](<https://devfeed.tech/tags/asynchronous-programming.md>), [call](<https://devfeed.tech/tags/call.md>), [client](<https://devfeed.tech/tags/client.md>), [data](<https://devfeed.tech/tags/data.md>), [data-science](<https://devfeed.tech/tags/data-science.md>), [editor-s-picks](<https://devfeed.tech/tags/editor-s-picks.md>), [first](<https://devfeed.tech/tags/first.md>), [free](<https://devfeed.tech/tags/free.md>), [google](<https://devfeed.tech/tags/google.md>), [google-adk](<https://devfeed.tech/tags/google-adk.md>), [human-in-the-loop](<https://devfeed.tech/tags/human-in-the-loop.md>), [open](<https://devfeed.tech/tags/open.md>), [post](<https://devfeed.tech/tags/post.md>), [programming](<https://devfeed.tech/tags/programming.md>), [resume](<https://devfeed.tech/tags/resume.md>), [rss](<https://devfeed.tech/tags/rss.md>), [science](<https://devfeed.tech/tags/science.md>), [webhooks](<https://devfeed.tech/tags/webhooks.md>), [works](<https://devfeed.tech/tags/works.md>)

### AI overview

Synchronous HTTP can leave a client blocked while a long-running agent works; the example RSS run held a connection open for 83 seconds. The article demonstrates an asynchronous pattern in Google ADK: return HTTP 202 with a session ID, run the task in the background, and send status or results to a webhook. For human approval pauses that must survive process restarts, it uses DatabaseSessionService with ResumabilityConfig, RequestInput interruptions, and interrupt IDs; on resume, ADK replays the workflow, reusing completed node results and running new paths. The initial asyncio.create_task approach is fragile: process or webhook failure can lose work or reports, and the article notes Cloud Run may throttle in-process background tasks after the response. Durable retries are deferred to a planned production follow-up using Cloud Tasks.

### Source excerpt

A synchronous call kept an RSS agent's connection open for 83 seconds; webhooks and 202 receipts free the client while the agent works. The post Google ADK Agents That Pause, Resume, and Wait Hours for a Human appeared first on Towards Data Science.

## How to Design MCP Servers That Leave Orchestration to the Host

DevFeed: [How to Design MCP Servers That Leave Orchestration to the Host](<https://devfeed.tech/articles/how-to-design-mcp-servers-that-leave-orchestration-to-the-host-137193.md>)

Original publisher: [Read original article](<https://towardsdatascience.com/how-to-design-mcp-servers-that-leave-orchestration-to-the-host/>)

Author: Priyansh Bhardwaj

Published: 2026-11-17T12:00:00Z

Content type: article

Language: en

Sources: [Towards Data Science](<https://devfeed.tech/sources/towards-data-science.md>)

Topics: [AI Architecture](<https://devfeed.tech/topics/ai-architecture.md>), [Orchestration](<https://devfeed.tech/topics/orchestration.md>), [API Governance](<https://devfeed.tech/topics/api-governance.md>), [Back end](<https://devfeed.tech/topics/backend.md>)

Tags: [adapter](<https://devfeed.tech/tags/adapter.md>), [agent-architecture](<https://devfeed.tech/tags/agent-architecture.md>), [agentic](<https://devfeed.tech/tags/agentic.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [ai-engineering](<https://devfeed.tech/tags/ai-engineering.md>), [api](<https://devfeed.tech/tags/api.md>), [audit](<https://devfeed.tech/tags/audit.md>), [design](<https://devfeed.tech/tags/design.md>), [generative-ai](<https://devfeed.tech/tags/generative-ai.md>), [mcp](<https://devfeed.tech/tags/mcp.md>), [mcp-servers](<https://devfeed.tech/tags/mcp-servers.md>), [model-context-protocol](<https://devfeed.tech/tags/model-context-protocol.md>), [orchestration](<https://devfeed.tech/tags/orchestration.md>), [planning](<https://devfeed.tech/tags/planning.md>), [prompt](<https://devfeed.tech/tags/prompt.md>), [python](<https://devfeed.tech/tags/python.md>), [python-sdk](<https://devfeed.tech/tags/python-sdk.md>), [resource](<https://devfeed.tech/tags/resource.md>), [sdk](<https://devfeed.tech/tags/sdk.md>), [server](<https://devfeed.tech/tags/server.md>), [software-architecture](<https://devfeed.tech/tags/software-architecture.md>), [state](<https://devfeed.tech/tags/state.md>), [tools](<https://devfeed.tech/tags/tools.md>), [v2](<https://devfeed.tech/tags/v2.md>), [workflow](<https://devfeed.tech/tags/workflow.md>)

### AI overview

Treat MCP as an integration boundary: servers expose domain capabilities, while hosts own planning, user interaction, workflow state, identity, deadlines and consequential-action policy. The Python SDK v2 catalogue example uses two narrow read-only tools, one JSON resource and one bounded prompt, with no planner or model client. It returns structured fields and limits dependency lookup to direct dependencies. The host can call only what the request needs, making the same server usable across different applications. The excerpt says MCP 2026-07-28 removed session identifiers and the initialization handshake, making the protocol core stateless; durable business workflows still require explicit identifiers and an owned store. Remote deployments still need authentication, action- and tenant-specific authorization, timeout and consistency policies, and careful failure handling. Writes additionally call for explicit approval, idempotency where retries are possible, deadlines and audit records. A specialised server workflow can be justified when it has a clear owner and permission model.

### Source excerpt

A Python SDK v2 catalogue server with two narrow tools, one resource, and one prompt shows how to keep planning, workflow state, and authorisation host-side. The post How to Design MCP Servers That Leave Orchestration to the Host appeared first on Towards Data Science.

## After 10 Heads in a Row, Small LLMs Bet Up to 82% on an 11th

DevFeed: [After 10 Heads in a Row, Small LLMs Bet Up to 82% on an 11th](<https://devfeed.tech/articles/after-10-heads-in-a-row-small-llms-bet-up-to-82-on-an-11th-135974.md>)

Original publisher: [Read original article](<https://towardsdatascience.com/after-10-heads-in-a-row-small-llms-bet-up-to-82-on-an-11th/>)

Author: Utkarsh Mangal

Published: 2026-11-16T16:30:00Z

Content type: article

Language: en

Sources: [Towards Data Science](<https://devfeed.tech/sources/towards-data-science.md>)

Topics: [AI Models](<https://devfeed.tech/topics/ai-models.md>), [Open Source Models & Datasets](<https://devfeed.tech/topics/open-source-models-datasets.md>), [post-training](<https://devfeed.tech/topics/post-training.md>)

Tags: [2](<https://devfeed.tech/tags/2.md>), [2016](<https://devfeed.tech/tags/2016.md>), [2023](<https://devfeed.tech/tags/2023.md>), [3](<https://devfeed.tech/tags/3.md>), [after](<https://devfeed.tech/tags/after.md>), [base](<https://devfeed.tech/tags/base.md>), [bias](<https://devfeed.tech/tags/bias.md>), [cognitive-bias](<https://devfeed.tech/tags/cognitive-bias.md>), [coin](<https://devfeed.tech/tags/coin.md>), [data](<https://devfeed.tech/tags/data.md>), [data-science](<https://devfeed.tech/tags/data-science.md>), [first](<https://devfeed.tech/tags/first.md>), [independent](<https://devfeed.tech/tags/independent.md>), [large-language-models](<https://devfeed.tech/tags/large-language-models.md>), [llm](<https://devfeed.tech/tags/llm.md>), [llms](<https://devfeed.tech/tags/llms.md>), [models](<https://devfeed.tech/tags/models.md>), [open](<https://devfeed.tech/tags/open.md>), [points](<https://devfeed.tech/tags/points.md>), [post](<https://devfeed.tech/tags/post.md>), [probability](<https://devfeed.tech/tags/probability.md>), [science](<https://devfeed.tech/tags/science.md>), [small](<https://devfeed.tech/tags/small.md>), [statistics](<https://devfeed.tech/tags/statistics.md>), [up](<https://devfeed.tech/tags/up.md>)

### AI overview

Across eight small open models, the probability assigned to another heads or tails rose with streak length; after ten or more identical flips it ranged from 0.60 to 0.82, though 0.50 is correct for a known fair coin. Explicitly stating that flips are fair and independent reduced the effect only slightly, and instruction tuning changed little overall. Early in sequences, models showed a modest bias toward alternation. They adapted to coins with 30% versus 70% repeat rates, but their change in expected repeat rate was only 0.09 to 0.22 against the true 0.40 difference. Stated answers could diverge from predictions: Qwen2.5-1.5B-Instruct favored "less than 50%" when asked, yet assigned 69% to continuation after a planted streak of twelve. Results are limited to small models, letter-encoded coin flips, and prompt-sensitive measurements; the proposed pattern-copying explanation was not tested.

### Source excerpt

Telling eight base and instruction-tuned open models that the coin is fair and the flips are independent trims their streak bias by only a few points. The post After 10 Heads in a Row, Small LLMs Bet Up to 82% on an 11th appeared first on Towards Data Science.

## Why Human in the Loop AI Oversight Fails and How to Fix It

DevFeed: [Why Human in the Loop AI Oversight Fails and How to Fix It](<https://devfeed.tech/articles/why-human-in-the-loop-ai-oversight-fails-and-how-to-fix-it-135977.md>)

Original publisher: [Read original article](<https://towardsdatascience.com/why-human-in-the-loop-ai-oversight-fails-and-how-to-fix-it/>)

Author: Corné Potgieter

Published: 2026-11-16T15:00:00Z

Content type: article

Language: en

Sources: [Towards Data Science](<https://devfeed.tech/sources/towards-data-science.md>)

Topics: [ai-governance](<https://devfeed.tech/topics/ai-governance.md>), [human review](<https://devfeed.tech/topics/human-review.md>), [Data Quality](<https://devfeed.tech/topics/data-quality.md>), [human in the loop agent](<https://devfeed.tech/topics/human-in-the-loop-agent.md>), [human computer interaction](<https://devfeed.tech/topics/human-computer-interaction.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [ai-governance](<https://devfeed.tech/tags/ai-governance.md>), [ai-oversight](<https://devfeed.tech/tags/ai-oversight.md>), [artificial-intelligence](<https://devfeed.tech/tags/artificial-intelligence.md>), [cognitive-bias](<https://devfeed.tech/tags/cognitive-bias.md>), [data-product](<https://devfeed.tech/tags/data-product.md>), [editor-s-picks](<https://devfeed.tech/tags/editor-s-picks.md>), [human-in-the-loop](<https://devfeed.tech/tags/human-in-the-loop.md>), [standards](<https://devfeed.tech/tags/standards.md>)

### AI overview

The article argues that routine human approval can become rubber stamping when reviewers face fluent AI output at scale and rely on quick intuition. It proposes structuring review with named data product owners (ODPS), output contracts covering schemas and quality rules (ODCS), and shared business definitions exchanged through Ossie. Automated checks and small owner-led feedback loops catch recurring issues; only exceptions that need judgment reach a human, with a specific question and compact context to prompt deliberate thought. In the PlayNest example, a cookie banner change made rejecting consent easier, reducing tracked visits and making the conversion rate appear to double. A contract check surfaces the change so an owner can update the relevant rule or definition. The article notes that Ossie provides formats for semantic exchange and an ontology, but does not determine business meaning, which remains under discussion.

### Source excerpt

Kahneman's System 1 and System 2, plus the ODPS, ODCS, and Ossie standards, show how to wake reviewers up only when a check fails. The post Why Human in the Loop AI Oversight Fails and How to Fix It appeared first on Towards Data Science.

## Local LLMs Handle Textbook RAG but Fail on New Research Papers

DevFeed: [Local LLMs Handle Textbook RAG but Fail on New Research Papers](<https://devfeed.tech/articles/local-llms-handle-textbook-rag-but-fail-on-new-research-papers-135976.md>)

Original publisher: [Read original article](<https://towardsdatascience.com/local-llms-handle-textbook-rag-but-fail-on-new-research-papers/>)

Author: Waleed Esmail

Published: 2026-11-16T12:00:00Z

Content type: article

Language: en

Sources: [Towards Data Science](<https://devfeed.tech/sources/towards-data-science.md>)

Topics: [Local AI](<https://devfeed.tech/topics/local-ai.md>), [generalization in machine learning](<https://devfeed.tech/topics/generalization-in-machine-learning.md>), [Hallucination detection](<https://devfeed.tech/topics/hallucination-detection.md>), [gpt-oss](<https://devfeed.tech/topics/gpt-oss.md>)

Tags: [large-language-model](<https://devfeed.tech/tags/large-language-model.md>), [large-language-models](<https://devfeed.tech/tags/large-language-models.md>), [llms](<https://devfeed.tech/tags/llms.md>), [local](<https://devfeed.tech/tags/local.md>), [local-llm-inference](<https://devfeed.tech/tags/local-llm-inference.md>), [local-llms](<https://devfeed.tech/tags/local-llms.md>), [papers](<https://devfeed.tech/tags/papers.md>), [rag](<https://devfeed.tech/tags/rag.md>), [rag-retrieval-augmented-generation](<https://devfeed.tech/tags/rag-retrieval-augmented-generation.md>)

### AI overview

In this evaluation, five setups scored 12-20 of 20 on textbook questions but only 0-9 of 15 on a paper published that year, despite identical retrieval and prompts. The author concludes that small local models handle familiar material better than new research, where they must rely on supplied passages. Grading answers alongside retrieved passages showed many errors came from missing evidence: figure descriptions raised figure-question results from 1/10 to 9/10, chapter summary trees raised overview results from 1/8 to 8/8, and neighboring chunks improved complete retrieval from 16/20 to 19/20. Quantization did not explain the gap. Qwen3.8-27B was close to Opus in the reported comparison at lower cost, but took about three minutes per answer. The author recommends writing gold questions first, separating familiar from new material, inspecting passages when grading, improving retrieval before changing models, and routing figures, web reading, and derivations to stronger models. Results are limited by a 35-question evaluation and grader drift.

### Source excerpt

Five setups, one blind grading pack: textbook scores ran 12 to 20 of 20, but a paper published this year dropped them to 0 to 9 of 15. The post Local LLMs Handle Textbook RAG but Fail on New Research Papers appeared first on Towards Data Science.

## A Practitioner's Guide to 6 Agentic GraphRAG Patterns

DevFeed: [A Practitioner's Guide to 6 Agentic GraphRAG Patterns](<https://devfeed.tech/articles/a-practitioner-s-guide-to-6-agentic-graphrag-patterns-134586.md>)

Original publisher: [Read original article](<https://towardsdatascience.com/a-practitioners-guide-to-6-agentic-graphrag-patterns/>)

Author: Partha Sarkar

Published: 2026-11-15T15:00:00Z

Content type: article

Language: en

Sources: [Towards Data Science](<https://devfeed.tech/sources/towards-data-science.md>)

Topics: [Agentic AI](<https://devfeed.tech/topics/what-is-agentic-ai.md>), [Multi-step AI processes](<https://devfeed.tech/topics/multi-step-ai-processes.md>), [agentic workflows](<https://devfeed.tech/topics/agentic-workflows.md>), [AI Context Management](<https://devfeed.tech/topics/ai-context-management.md>), [Retrieval Augmented Generation (RAG)](<https://devfeed.tech/topics/retrieval-augmented-generation-rag.md>), [knowledge-engineering](<https://devfeed.tech/topics/knowledge-engineering.md>), [Graph Mining](<https://devfeed.tech/topics/graph-mining.md>), [Multi Agent Systems](<https://devfeed.tech/topics/multi-agent-systems.md>)

Tags: [abstraction](<https://devfeed.tech/tags/abstraction.md>), [accelerated](<https://devfeed.tech/tags/accelerated.md>), [access-control](<https://devfeed.tech/tags/access-control.md>), [act](<https://devfeed.tech/tags/act.md>), [actions](<https://devfeed.tech/tags/actions.md>), [adaptive](<https://devfeed.tech/tags/adaptive.md>), [advanced](<https://devfeed.tech/tags/advanced.md>), [agentic](<https://devfeed.tech/tags/agentic.md>), [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [agentic-graphrag](<https://devfeed.tech/tags/agentic-graphrag.md>), [agents](<https://devfeed.tech/tags/agents.md>), [autonomous](<https://devfeed.tech/tags/autonomous.md>), [autonomous-agents](<https://devfeed.tech/tags/autonomous-agents.md>), [context](<https://devfeed.tech/tags/context.md>), [context-graphs](<https://devfeed.tech/tags/context-graphs.md>), [graphrag](<https://devfeed.tech/tags/graphrag.md>), [guide](<https://devfeed.tech/tags/guide.md>), [large-language-models](<https://devfeed.tech/tags/large-language-models.md>), [llm](<https://devfeed.tech/tags/llm.md>), [operational](<https://devfeed.tech/tags/operational.md>), [patterns](<https://devfeed.tech/tags/patterns.md>), [practitioner](<https://devfeed.tech/tags/practitioner.md>), [rag](<https://devfeed.tech/tags/rag.md>), [user-intent](<https://devfeed.tech/tags/user-intent.md>)

### AI overview

The guide presents Agentic GraphRAG as a knowledge graph that agents navigate step by step and update with operational history, decisions, and user context. It outlines six patterns: operational memory for precedents and exceptions; execution memory for reusing successful tool paths and avoiding known failures; user context graphs for query-specific cross-session preferences; self-enriching topology graphs for incident matching, responder routing, and blast-radius analysis; hierarchical summaries for drilling from broad questions to atomic facts; and ReAct navigation for verified multi-hop retrieval. The proposed benefits include more relevant context, fewer repeated tool failures, and better grounded investigations. Trade-offs include schema and lifecycle governance, access controls, fresh incident and summary pipelines, and added latency and safeguards for iterative traversal. The article describes architectures and illustrative scenarios; it does not establish independent performance evidence for its quantified benefit claims.

### Source excerpt

How autonomous agents turn static knowledge graphs into living Context Graphs that remember decisions, user intent, and operational history The post A Practitioner's Guide to 6 Agentic GraphRAG Patterns appeared first on Towards Data Science.

## Computing Effective Dimension for Quantum ML Models in PyTorch

DevFeed: [Computing Effective Dimension for Quantum ML Models in PyTorch](<https://devfeed.tech/articles/computing-effective-dimension-for-quantum-ml-models-in-pytorch-134418.md>)

Original publisher: [Read original article](<https://towardsdatascience.com/computing-effective-dimension-for-quantum-ml-models-in-pytorch/>)

Author: Adrian Panasiewicz

Published: 2026-11-15T13:00:00Z

Content type: tutorial

Language: en

Sources: [Towards Data Science](<https://devfeed.tech/sources/towards-data-science.md>)

Topics: [PyTorch](<https://devfeed.tech/topics/pytorch.md>), [Quantum Computing](<https://devfeed.tech/topics/quantum-computing.md>), [AI Models](<https://devfeed.tech/topics/ai-models.md>)

Tags: [code](<https://devfeed.tech/tags/code.md>), [computing](<https://devfeed.tech/tags/computing.md>), [data](<https://devfeed.tech/tags/data.md>), [deep-learning](<https://devfeed.tech/tags/deep-learning.md>), [dimension](<https://devfeed.tech/tags/dimension.md>), [editor-s-picks](<https://devfeed.tech/tags/editor-s-picks.md>), [effective](<https://devfeed.tech/tags/effective.md>), [experiment](<https://devfeed.tech/tags/experiment.md>), [first](<https://devfeed.tech/tags/first.md>), [global](<https://devfeed.tech/tags/global.md>), [local](<https://devfeed.tech/tags/local.md>), [log](<https://devfeed.tech/tags/log.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [ml](<https://devfeed.tech/tags/ml.md>), [ml-models](<https://devfeed.tech/tags/ml-models.md>), [models](<https://devfeed.tech/tags/models.md>), [pytorch](<https://devfeed.tech/tags/pytorch.md>), [quantum](<https://devfeed.tech/tags/quantum.md>), [quantum-computing](<https://devfeed.tech/tags/quantum-computing.md>), [quantum-machine-learning](<https://devfeed.tech/tags/quantum-machine-learning.md>), [sampling](<https://devfeed.tech/tags/sampling.md>), [science](<https://devfeed.tech/tags/science.md>), [series](<https://devfeed.tech/tags/series.md>), [space](<https://devfeed.tech/tags/space.md>), [statistics](<https://devfeed.tech/tags/statistics.md>), [time](<https://devfeed.tech/tags/time.md>)

### AI overview

The article shows how to estimate effective dimension for quantum ML models in PyTorch using an empirical Fisher matrix and Monte Carlo parameter sampling. It distinguishes global sampling across quantum rotation angles in [0, 2π] from local sampling inside an ε-ball around trained weights; for uniform ball coverage, it samples a normalized Gaussian direction and scales the radius by εq^(1/k). It recommends skipping outcomes below a probability threshold and computing log-determinants with `torch.linalg.slogdet` and sample averages with `torch.logsumexp` to avoid numerical instability. In a PennyLane/PyTorch Mackey-Glass experiment, added circuit depth initially improved test loss and effective dimension, while deeper models showed diminishing gains and redundant parameter directions. The method is computationally expensive: runtime scales with parameter samples times data samples, and the article estimates several days for training and evaluation on high-end computers.

### Source excerpt

Local and global variants, Fisher estimation, epsilon-ball sampling, and log-space stability, with PennyLane code and a time-series experiment. The post Computing Effective Dimension for Quantum ML Models in PyTorch appeared first on Towards Data Science.

## How to Build a Markov Chain in Python With NumPy

DevFeed: [How to Build a Markov Chain in Python With NumPy](<https://devfeed.tech/articles/how-to-build-a-markov-chain-in-python-with-numpy-134066.md>)

Original publisher: [Read original article](<https://towardsdatascience.com/how-to-build-a-markov-chain-in-python-with-numpy/>)

Author: Ananya Bhattacharyya

Published: 2026-11-14T15:00:00Z

Content type: article

Language: en

Sources: [Towards Data Science](<https://devfeed.tech/sources/towards-data-science.md>)

Topics: [Python](<https://devfeed.tech/topics/python.md>)

Tags: [analytics](<https://devfeed.tech/tags/analytics.md>), [credit-risk](<https://devfeed.tech/tags/credit-risk.md>), [customer-analytics](<https://devfeed.tech/tags/customer-analytics.md>), [data-science](<https://devfeed.tech/tags/data-science.md>), [distributions](<https://devfeed.tech/tags/distributions.md>), [editor-s-picks](<https://devfeed.tech/tags/editor-s-picks.md>), [example](<https://devfeed.tech/tags/example.md>), [guide](<https://devfeed.tech/tags/guide.md>), [markov](<https://devfeed.tech/tags/markov.md>), [markov-chain](<https://devfeed.tech/tags/markov-chain.md>), [markov-chains](<https://devfeed.tech/tags/markov-chains.md>), [numpy](<https://devfeed.tech/tags/numpy.md>), [python](<https://devfeed.tech/tags/python.md>), [statistics](<https://devfeed.tech/tags/statistics.md>), [writing](<https://devfeed.tech/tags/writing.md>)

### AI overview

The guide builds a three-state writing model--Drafting, Revising, and Distracted--to show how a row-stochastic transition matrix yields future probabilities through matrix powers. Starting in Drafting, it calculates a 37% chance of Revising two days later; under the example's assumptions, probabilities approach a stationary distribution of 40% Drafting, 40% Revising, and 20% Distracted. It describes using NumPy for these calculations and Matplotlib to plot convergence, and simulating a single random 30-day path to distinguish an individual trajectory from the probability distribution. The result depends on the state definitions and assumptions: the model is memoryless and time-homogeneous, and convergence requires conditions such as irreducibility and aperiodicity. If yesterday's state affects tomorrow, the guide suggests a second-order model or richer states; adding an absorbing Published state changes the long-run behavior.

### Source excerpt

A step-by-step guide to transition matrices, the Chapman-Kolmogorov equation, and stationary distributions, built around a three-state writing example. The post How to Build a Markov Chain in Python With NumPy appeared first on Towards Data Science.

## Data From 100+ AI Agents Shows They Fail by Silent Drift, Not Errors

DevFeed: [Data From 100+ AI Agents Shows They Fail by Silent Drift, Not Errors](<https://devfeed.tech/articles/data-from-100-ai-agents-shows-they-fail-by-silent-drift-not-errors-134065.md>)

Original publisher: [Read original article](<https://towardsdatascience.com/data-from-100-ai-agents-shows-they-fail-by-silent-drift-not-errors/>)

Author: Barr Moses

Published: 2026-11-14T13:00:00Z

Content type: article

Language: en

Sources: [Towards Data Science](<https://devfeed.tech/sources/towards-data-science.md>)

Topics: [LLM observability](<https://devfeed.tech/topics/llm-observability.md>), [ai security](<https://devfeed.tech/topics/ai-security.md>)

Tags: [2](<https://devfeed.tech/tags/2.md>), [about](<https://devfeed.tech/tags/about.md>), [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [agentops](<https://devfeed.tech/tags/agentops.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [ai-evaluation](<https://devfeed.tech/tags/ai-evaluation.md>), [ai-monitoring](<https://devfeed.tech/tags/ai-monitoring.md>), [alerts](<https://devfeed.tech/tags/alerts.md>), [cost](<https://devfeed.tech/tags/cost.md>), [data](<https://devfeed.tech/tags/data.md>), [data-science](<https://devfeed.tech/tags/data-science.md>), [editor-s-picks](<https://devfeed.tech/tags/editor-s-picks.md>), [errors](<https://devfeed.tech/tags/errors.md>), [fail](<https://devfeed.tech/tags/fail.md>), [first](<https://devfeed.tech/tags/first.md>), [latency](<https://devfeed.tech/tags/latency.md>), [monitoring](<https://devfeed.tech/tags/monitoring.md>), [post](<https://devfeed.tech/tags/post.md>), [quality](<https://devfeed.tech/tags/quality.md>), [science](<https://devfeed.tech/tags/science.md>), [shows](<https://devfeed.tech/tags/shows.md>), [token](<https://devfeed.tech/tags/token.md>), [with](<https://devfeed.tech/tags/with.md>)

### AI overview

In alert data from more than 100 production agents monitored across dozens of companies from April 9 to November 2, 2026, quality and token alerts were each among the first signals for 38% of agents, while errors were first for about 1%. Quality, token, and latency checks often fired within days; error alerts were less common and took a median of 42 days when they occurred. Alerts frequently recurred, and token spikes often coincided with quality, latency, error, or loop warnings. The analysis suggests monitoring quality, cost, and speed early, inspecting individual tool calls for loops and token use, and explicitly adding governance and safety checks. The findings describe alerts, not confirmed failures or causes; coverage was uneven, and the data came from customers of one monitoring platform, with many monitors configured by the vendor.

### Source excerpt

Quality and token alerts were early signals for 38% of agents and errors for about 1%, so monitoring should start with quality, cost, and latency. The post Data From 100+ AI Agents Shows They Fail by Silent Drift, Not Errors appeared first on Towards Data Science.

## Keyword Search Tied an 8B Embedding Model at Picking Agent Tools

DevFeed: [Keyword Search Tied an 8B Embedding Model at Picking Agent Tools](<https://devfeed.tech/articles/keyword-search-tied-an-8b-embedding-model-at-picking-agent-tools-132593.md>)

Original publisher: [Read original article](<https://towardsdatascience.com/keyword-search-tied-an-8b-embedding-model-at-picking-agent-tools/>)

Author: Ankur Tomar

Published: 2026-11-13T16:30:00Z

Content type: article

Language: en

Sources: [Towards Data Science](<https://devfeed.tech/sources/towards-data-science.md>)

Topics: [AI-driven information retrieval](<https://devfeed.tech/topics/ai-driven-information-retrieval.md>), [Embeddings](<https://devfeed.tech/topics/embeddings.md>), [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [context window](<https://devfeed.tech/topics/context-window.md>), [Large Language Model](<https://devfeed.tech/topics/llm.md>), [LLM optimization](<https://devfeed.tech/topics/llm-optimization.md>), [Local AI](<https://devfeed.tech/topics/local-ai.md>), [hybrid-search](<https://devfeed.tech/topics/hybrid-search.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agent-tools](<https://devfeed.tech/tags/agent-tools.md>), [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [benchmarking](<https://devfeed.tech/tags/benchmarking.md>), [bm25](<https://devfeed.tech/tags/bm25.md>), [cache](<https://devfeed.tech/tags/cache.md>), [caching](<https://devfeed.tech/tags/caching.md>), [claude](<https://devfeed.tech/tags/claude.md>), [claude-opus](<https://devfeed.tech/tags/claude-opus.md>), [claude-opus-5](<https://devfeed.tech/tags/claude-opus-5.md>), [code](<https://devfeed.tech/tags/code.md>), [code-style](<https://devfeed.tech/tags/code-style.md>), [cpu](<https://devfeed.tech/tags/cpu.md>), [embedding](<https://devfeed.tech/tags/embedding.md>), [embedding-model](<https://devfeed.tech/tags/embedding-model.md>), [information-retrieval](<https://devfeed.tech/tags/information-retrieval.md>), [llm-agents](<https://devfeed.tech/tags/llm-agents.md>), [model](<https://devfeed.tech/tags/model.md>), [rag](<https://devfeed.tech/tags/rag.md>), [search](<https://devfeed.tech/tags/search.md>), [tool](<https://devfeed.tech/tags/tool.md>), [tool-calling](<https://devfeed.tech/tags/tool-calling.md>), [tools](<https://devfeed.tech/tags/tools.md>)

### AI overview

In a 200-question ToolRet experiment using up to 3,794 code-category tools, every retrieval method lost accuracy as the catalogue grew. At the largest size, BM25, BGE-large, Qwen3-Embedding-8B, and a reranking pipeline scored similarly; BM25 was about 40 times faster than BGE-large and needed no GPU. BM25 tokenization must split code-style tool names on underscores and capitals. Using full tool documentation improved BM25 and BGE-large by roughly 0.02-0.04 nDCG; BM25 with full documentation scored 0.533 at about a millisecond per question, though Qwen3-8B and reranking were not tested with full docs. Reranking increased Recall@10 but usually lowered nDCG, suggesting it found more relevant tools without ordering them as well. Prompting with the entire catalogue was too slow to evaluate at scale and exceeded Gemma 3's context window beyond about 2,500 trimmed tool descriptions. Even the best setup included every needed tool in its top ten on only 65% of tasks. Results are limited by 200 questions and a code-focused category that may favor keyword matching.

### Source excerpt

On 200 ToolRet questions and up to 3,794 tools, BM25 ran about 40x faster than BGE-large on just a CPU, and full tool docs made it the best setup tested. The post Keyword Search Tied an 8B Embedding Model at Picking Agent Tools appeared first on Towards Data Science.

## How Jev Makes Decisions About Images, With 4 Python Examples

DevFeed: [How Jev Makes Decisions About Images, With 4 Python Examples](<https://devfeed.tech/articles/how-jev-makes-decisions-about-images-with-4-python-examples-132592.md>)

Original publisher: [Read original article](<https://towardsdatascience.com/how-jev-makes-decisions-about-images-with-4-python-examples/>)

Author: Thomas Reid

Published: 2026-11-13T15:00:00Z

Content type: tutorial

Language: en

Sources: [Towards Data Science](<https://devfeed.tech/sources/towards-data-science.md>)

Topics: [Image](<https://devfeed.tech/topics/image.md>), [Image processing](<https://devfeed.tech/topics/image-processing.md>), [AI Integration Strategies](<https://devfeed.tech/topics/ai-integration-strategies.md>)

Tags: [2](<https://devfeed.tech/tags/2.md>), [3](<https://devfeed.tech/tags/3.md>), [about](<https://devfeed.tech/tags/about.md>), [act](<https://devfeed.tech/tags/act.md>), [after](<https://devfeed.tech/tags/after.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-applications](<https://devfeed.tech/tags/ai-applications.md>), [ai-models](<https://devfeed.tech/tags/ai-models.md>), [all](<https://devfeed.tech/tags/all.md>), [car](<https://devfeed.tech/tags/car.md>), [code](<https://devfeed.tech/tags/code.md>), [editor-s-picks](<https://devfeed.tech/tags/editor-s-picks.md>), [examples](<https://devfeed.tech/tags/examples.md>), [fabric](<https://devfeed.tech/tags/fabric.md>), [first](<https://devfeed.tech/tags/first.md>), [image-classification](<https://devfeed.tech/tags/image-classification.md>), [image-processing](<https://devfeed.tech/tags/image-processing.md>), [images](<https://devfeed.tech/tags/images.md>), [jev](<https://devfeed.tech/tags/jev.md>), [practical](<https://devfeed.tech/tags/practical.md>), [product](<https://devfeed.tech/tags/product.md>), [python](<https://devfeed.tech/tags/python.md>), [reference](<https://devfeed.tech/tags/reference.md>), [shipping](<https://devfeed.tech/tags/shipping.md>), [with](<https://devfeed.tech/tags/with.md>)

### AI overview

JEV's image API turns visual checks into structured decisions that software can act on. The excerpt demonstrates classifying fabric against reference photos, scoring product sharpness from 0-3, confirming and cropping a shipping label, and comparing car-door images to locate a scratch. The new Box primitive returns image-relative bounding-box edges; the shipping-label example first checks presence, then crops only when label probability is at least 90% and box confidence at least 80%. Other examples use confidence or probability thresholds to pass, request a retake, or flag manual review, and the author notes thresholds may need adjustment. Images must be JPEG or PNG, have dimensions from 280 to 1568 pixels, and be no larger than 16 MiB each; request bodies are limited to 32 MiB. The excerpt describes example code and outputs but does not establish general accuracy.

### Source excerpt

Practical code for classifying fabric against reference photos, scoring product sharpness, cropping shipping labels, and spotting subtle car damage. The post How Jev Makes Decisions About Images, With 4 Python Examples appeared first on Towards Data Science.

## What AI Agents Speed Up in PhD Research, and What They Can't

DevFeed: [What AI Agents Speed Up in PhD Research, and What They Can't](<https://devfeed.tech/articles/what-ai-agents-speed-up-in-phd-research-and-what-they-can-t-131790.md>)

Original publisher: [Read original article](<https://towardsdatascience.com/what-ai-agents-speed-up-in-phd-research-and-what-they-cant/>)

Author: Fiona Victoria

Published: 2026-11-13T13:30:00Z

Content type: opinion

Language: en

Sources: [Towards Data Science](<https://devfeed.tech/sources/towards-data-science.md>)

Topics: [Tool](<https://devfeed.tech/topics/tool.md>), [AI-assisted coding](<https://devfeed.tech/topics/ai-assisted-coding.md>), [Boilerplate](<https://devfeed.tech/topics/boilerplate.md>), [Koans](<https://devfeed.tech/topics/koans.md>)

Tags: [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-adoption](<https://devfeed.tech/tags/ai-adoption.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [artificial-intelligence](<https://devfeed.tech/tags/artificial-intelligence.md>), [career-advice](<https://devfeed.tech/tags/career-advice.md>), [change](<https://devfeed.tech/tags/change.md>), [code](<https://devfeed.tech/tags/code.md>), [data](<https://devfeed.tech/tags/data.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [faster](<https://devfeed.tech/tags/faster.md>), [higher-education](<https://devfeed.tech/tags/higher-education.md>), [labs](<https://devfeed.tech/tags/labs.md>), [ml](<https://devfeed.tech/tags/ml.md>), [phd-research](<https://devfeed.tech/tags/phd-research.md>), [predictions](<https://devfeed.tech/tags/predictions.md>), [research](<https://devfeed.tech/tags/research.md>), [science](<https://devfeed.tech/tags/science.md>), [speed](<https://devfeed.tech/tags/speed.md>), [student](<https://devfeed.tech/tags/student.md>)

### AI overview

AI agents speed up research execution: they help with boilerplate code, debugging unfamiliar tools, running more experiments, literature searches, and drafting. The author says this makes researchers' work shift toward choosing worthwhile questions and checking whether results are sound; faster, correct code can still produce a wrong conclusion when labels or observations misrepresent reality. In ecology, species may go undetected because of weather or behavior, so understanding how data was collected can motivate occupancy models that distinguish presence from detectability. The author predicts researchers will act more like team leads, domain knowledge will matter more, and labs will become more interdisciplinary. Human judgment, curiosity, expert trust, and accountability remain central.

### Source excerpt

A PhD student in ML and ecology on why faster code, experiments, and drafts don't fix data with a backstory, and three predictions for how labs will change. The post What AI Agents Speed Up in PhD Research, and What They Can't appeared first on Towards Data Science.

## How to Check Local LLM Output with Ollama, Gemma 4, and Nimble

DevFeed: [How to Check Local LLM Output with Ollama, Gemma 4, and Nimble](<https://devfeed.tech/articles/how-to-check-local-llm-output-with-ollama-gemma-4-and-nimble-131789.md>)

Original publisher: [Read original article](<https://towardsdatascience.com/how-to-check-local-llm-output-with-ollama-gemma-4-and-nimble/>)

Author: Shuai Guo

Published: 2026-11-13T12:00:00Z

Content type: tutorial

Language: en

Sources: [Towards Data Science](<https://devfeed.tech/sources/towards-data-science.md>)

Topics: [Ollama](<https://devfeed.tech/topics/ollama.md>), [Local AI](<https://devfeed.tech/topics/local-ai.md>), [Large Language Model](<https://devfeed.tech/topics/llm.md>), [gemma](<https://devfeed.tech/topics/gemma.md>), [human review](<https://devfeed.tech/topics/human-review.md>)

Tags: [editor-s-picks](<https://devfeed.tech/tags/editor-s-picks.md>), [gemma](<https://devfeed.tech/tags/gemma.md>), [gemma-4](<https://devfeed.tech/tags/gemma-4.md>), [how-to](<https://devfeed.tech/tags/how-to.md>), [llm](<https://devfeed.tech/tags/llm.md>), [llm-agents](<https://devfeed.tech/tags/llm-agents.md>), [llm-applications](<https://devfeed.tech/tags/llm-applications.md>), [local](<https://devfeed.tech/tags/local.md>), [local-llm](<https://devfeed.tech/tags/local-llm.md>), [nimble](<https://devfeed.tech/tags/nimble.md>), [ollama](<https://devfeed.tech/tags/ollama.md>), [slm](<https://devfeed.tech/tags/slm.md>)

### AI overview

The article builds a local workflow in Ollama where Gemma 4 E2B drafts three attendee replies from team messages and Nimble, Bespoke Labs' 9B decision model, reviews them as ACCEPT, REVISE_LOCALLY, or HUMAN_REVIEW. Python keeps accepted answers, retries only answers Nimble marks locally correctable using the same context and requirement but one question at a time, and flags cases lacking enough information for a person. In the example, Nimble catches Gemma's mistaken claim that the attendee reminder was sent; the focused retry correctly says it was cancelled pending speaker-list sign-off. The setup requires Ollama 0.35.0 or later for Nimble's decision endpoint. Both models run via Ollama's local API, keeping the supplied context on the machine. The article cautions that results depend on models, prompts, and task, and recommends testing on task-specific examples.

### Source excerpt

Gemma 4 E2B drafts the replies, Bespoke Labs' Nimble 9B picks accept, revise, or human review, and only flagged questions get a focused retry. The post How to Check Local LLM Output with Ollama, Gemma 4, and Nimble appeared first on Towards Data Science.

## RAG Retrieval Lost 40 Points When Distracting Passages Grew to 2.68M

DevFeed: [RAG Retrieval Lost 40 Points When Distracting Passages Grew to 2.68M](<https://devfeed.tech/articles/rag-retrieval-lost-40-points-when-distracting-passages-grew-to-2-68m-130836.md>)

Original publisher: [Read original article](<https://towardsdatascience.com/rag-retrieval-lost-40-points-when-distracting-passages-grew-to-268m/>)

Author: Abdullahi Dattijo

Published: 2026-11-12T16:30:00Z

Content type: article

Language: en

Sources: [Towards Data Science](<https://devfeed.tech/sources/towards-data-science.md>)

Topics: [Retrieval Augmented Generation (RAG)](<https://devfeed.tech/topics/retrieval-augmented-generation-rag.md>), [information retrieval](<https://devfeed.tech/topics/information-retrieval.md>), [search-engines](<https://devfeed.tech/topics/search-engines.md>), [Embeddings](<https://devfeed.tech/topics/embeddings.md>), [context window](<https://devfeed.tech/topics/context-window.md>), [Vectors & LLMs](<https://devfeed.tech/topics/vectors-llms.md>)

Tags: [bm25](<https://devfeed.tech/tags/bm25.md>), [context](<https://devfeed.tech/tags/context.md>), [embeddings](<https://devfeed.tech/tags/embeddings.md>), [information-retrieval](<https://devfeed.tech/tags/information-retrieval.md>), [large-language-models](<https://devfeed.tech/tags/large-language-models.md>), [llm](<https://devfeed.tech/tags/llm.md>), [llm-evaluation](<https://devfeed.tech/tags/llm-evaluation.md>), [position](<https://devfeed.tech/tags/position.md>), [rag](<https://devfeed.tech/tags/rag.md>), [retrieval](<https://devfeed.tech/tags/retrieval.md>), [search](<https://devfeed.tech/tags/search.md>), [test](<https://devfeed.tech/tags/test.md>)

### AI overview

BM25 retrieval was nearly unchanged as a collection grew to 1,801 arXiv papers, but on 2.68 million Wikipedia passages its success at 10 fell about 40 points versus a 10,000-passage baseline; the tests differed in several ways, so the author cannot identify which difference drove the gap. In a smaller Wikipedia test, adding 20 unsupported passages from the answer's own page reduced retrieval success from 90-92% to 81-84%. The added passages often ranked highly because BM25 doubled title-match weight, not because they were semantically similar; embeddings were not tested there. With Qwen3 4B, answers scored worse when the judged passage was lower or absent in the retrieved context, and placing it first made scores nearly equal across conditions. The author recommends testing with questions written by people unfamiliar with the documents and measuring the answer passage's position. Results are limited by positive-only relevance labels, unusually high judged-passage density in the smallest Wikipedia collection, and a single reviewer for the same-page labels.

### Source excerpt

A reproducible test of BM25, embeddings, and parent documents shows why author-written evals overrate search, and why answer position in context matters. The post RAG Retrieval Lost 40 Points When Distracting Passages Grew to 2.68M appeared first on Towards Data Science.

## I Mutation-Tested 50 Python Libraries. Tests in the Young Ones Missed 41% of Planted Faults.

DevFeed: [I Mutation-Tested 50 Python Libraries. Tests in the Young Ones Missed 41% of Planted Faults.](<https://devfeed.tech/articles/i-mutation-tested-50-python-libraries-tests-in-the-young-ones-missed-41-of-planted-faults-130835.md>)

Original publisher: [Read original article](<https://towardsdatascience.com/i-mutation-tested-50-python-libraries-tests-in-the-young-ones-missed-41-of-planted-faults/>)

Author: Gal Arav

Published: 2026-11-12T15:00:00Z

Content type: article

Language: en

Sources: [Towards Data Science](<https://devfeed.tech/sources/towards-data-science.md>)

Topics: [mutation-testing](<https://devfeed.tech/topics/mutation-testing.md>), [Test automation](<https://devfeed.tech/topics/test-automation.md>), [Pytest](<https://devfeed.tech/topics/pytest.md>), [Python](<https://devfeed.tech/topics/python.md>)

Tags: [abstract](<https://devfeed.tech/tags/abstract.md>), [acceptance](<https://devfeed.tech/tags/acceptance.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agent-skill](<https://devfeed.tech/tags/agent-skill.md>), [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [ai-assistant](<https://devfeed.tech/tags/ai-assistant.md>), [ai-development](<https://devfeed.tech/tags/ai-development.md>), [ai-testing](<https://devfeed.tech/tags/ai-testing.md>), [bug](<https://devfeed.tech/tags/bug.md>), [data-loss](<https://devfeed.tech/tags/data-loss.md>), [editor-s-picks](<https://devfeed.tech/tags/editor-s-picks.md>), [libraries](<https://devfeed.tech/tags/libraries.md>), [llm-engineering](<https://devfeed.tech/tags/llm-engineering.md>), [mutation](<https://devfeed.tech/tags/mutation.md>), [python](<https://devfeed.tech/tags/python.md>), [python-libraries](<https://devfeed.tech/tags/python-libraries.md>), [quality-assessment](<https://devfeed.tech/tags/quality-assessment.md>), [readme](<https://devfeed.tech/tags/readme.md>), [software-testing](<https://devfeed.tech/tags/software-testing.md>), [study](<https://devfeed.tech/tags/study.md>), [tests](<https://devfeed.tech/tags/tests.md>), [windows](<https://devfeed.tech/tags/windows.md>)

### AI overview

Mutation testing of 50 active Python libraries found that tests missed 41% of planted faults in 30 young libraries and 17% in 20 mature libraries. Across 2,312 planted faults, the study also identified four real library problems, confirmed by hand and reported to maintainers. A missed mutant is only a lead: some changes are unobservable, and the method cannot reveal behavior absent from the code. The article's workflow plants faults, runs the existing suite, and turns misses into questions about intended behavior. Documentation and signatures can help an LLM draft answers, but coverage was limited (about 1 in 4 questions for mature libraries, 1 in 15 for young ones), and checked drafts were often wrong, especially for young libraries. The author recommends human review of answers before generating tests. In qikly, step 1 is free; steps 2 and 3 use Gemini, Anthropic, or OpenAI APIs and require a key. A Skill can convey practices to an assistant, while the local MCP server withholds acceptance criteria from the coding agent.

### Source excerpt

Across 2,312 planted faults the study also surfaced four real library problems, from a Windows data-loss bug to README examples that do not work as written. The post I Mutation-Tested 50 Python Libraries. Tests in the Young Ones Missed 41% of Planted Faults. appeared first on Towards Data Science.

[Next page](<https://devfeed.tech/sources/towards-data-science.md?cursor=WyIyMDI2LTExLTEyVDE1OjAwOjAwKzAwOjAwIiwgImQ1OWVlNzI3LWY2ZjAtNDFiNy1iNjVkLTkyNTZkNWU3MGE0YSJd>)