# Data Science Published articles for Data Science. This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles. ## Beyond the model: Engineering AI infra with scientific judgement DevFeed: [Beyond the model: Engineering AI infra with scientific judgement]() Original publisher: [Read original article]() Author: AirbnbEng Published: 2026-09-15T17:06:18Z Content type: article Language: en Sources: [The Airbnb Tech Blog - Medium]() Topics: [Agent Harness](), [AI Agent](), [Data Science](), [data](), [datasets](), [Large Language Model]() Tags: [agent-harness](), [ai](), [ai-agent](), [data](), [data-science](), [datasets](), [engineering](), [llms](), [quality](), [science](), [trust]() ### AI overview Airbnb describes an agent harness for data science that embeds scientific methodology around an AI model. The system guides agents through framing questions, selecting evidence, and recording decisions so unstructured-data investigations can be reproduced, audited, challenged, and extended across languages, geographies, and LLM-based products. ### Source excerpt How Airbnb's agent harness transforms unstructured data exploration by encoding scientific methodology into scalable, reproducible, and audit-ready infrastructure. By: Wren Dougherty Ask a coding agent to analyze 100,000 customer support conversations and within minutes you'll have a polished taxonomy, precise prevalence numbers, and an executive-ready summary. What you can't see is the investigation that produced them: the methods it chose, the evidence it weighed, how much to trust it, or whether a second request would agree. All that reaches you is the polish. The model is undeniably intelligent, but intelligence without methodology is not science. LLMs certainly make for confident scientists, but we need them to be responsible ones. Smarter models help, but intelligence has never been the whole of science, in people or in machines. The method is as much the product as the answer. That is the idea behind the agent harness we built for data science: the methodology itself, built as infrastructure around the model. It governs how an AI agent operates, from framing a question to selecting evidence to recording decisions, so results can be reproduced, audited, and challenged, and the method shared, inspected, and built on. The challenge of unstructured data exploration In 2025, Airbnb was preparing to launch an AI customer service assistant. Before it could ship, we needed to understand exactly what kinds of situations it would face in the real world. That included rare events that could be risky for AI to interact with, and involved examining their taxonomy and prevalence to create the datasets that would help us build a more responsible product. The investigative work to do this was rigorous, but the process was deeply artisanal. Months of high-touch iteration went into each investigation, from finding the right data, reviewing samples with experts, and generating representative datasets, and the method was manually curated across notebooks, tables, docs, and indiv ## Scaling Federated Learning Across Docker, Kubernetes, and Slurm with NVIDIA FLARE DevFeed: [Scaling Federated Learning Across Docker, Kubernetes, and Slurm with NVIDIA FLARE]() Original publisher: [Read original article]() Author: Elizabeth Goodman Published: 2026-09-15T15:00:00Z Content type: article Language: en Sources: [NVIDIA Developer](), [NVIDIA Technical Blog]() Topics: [Federated Learning](), [Docker](), [Kubernetes](), [Nvidia](), [Server](), [dataset](), [Deployment]() Tags: [architecture](), [compute](), [configuration](), [container](), [data-analytics-processing](), [data-science](), [dataset](), [deployment](), [docker](), [docker-container](), [federated-learning](), [gpu](), [job](), [kubernetes](), [nvidia](), [nvidia-flare](), [server]() ### AI overview This article explains how NVIDIA FLARE scales federated learning across sites with different infrastructure, including Docker, Kubernetes, and Slurm. Its two-layer architecture separates persistent federation services from on-demand job execution, while allowing each site to retain local control over compute, data, secrets, and scheduling. ### Source excerpt Federated learning (FL) projects often begin with a straightforward setup: one server, a few clients, and one dataset at each site. As those projects grow, the... ## High-Throughput Structure Prediction with BioNeMo Inference Runtime DevFeed: [High-Throughput Structure Prediction with BioNeMo Inference Runtime]() Original publisher: [Read original article]() Author: Elizabeth Goodman Published: 2026-09-10T15:00:00Z Content type: tutorial Language: en Sources: [NVIDIA Developer](), [NVIDIA Technical Blog]() Topics: [AI, ML & Data Engineering]() Tags: [bionemo](), [cuda](), [cuda-graphs](), [data-center-cloud](), [data-science](), [drug-discovery](), [gpu](), [healthcare-life-sciences](), [how-to](), [hpc-scientific-computing](), [inference](), [integration](), [kernels](), [model](), [models](), [node](), [nvidia](), [python](), [pytorch](), [resource](), [scale](), [simulation-modeling-design](), [tokenization](), [torch](), [tutorial](), [workflow]() ### AI overview A tutorial on using NVIDIA BioNeMo Inference Runtime to accelerate biomolecular structure-prediction models on GPUs. It covers the end-to-end Boltz2 workflow, PyTorch integration, input requirements, and Ray-based single-node throughput scaling. ### Source excerpt Biomolecular structure prediction is now often run at proteome scale, where the goal is to move an entire worklist through the pipeline efficiently. NVIDIA... ## Why Spotify Is Not Using Bayesian A/B Testing DevFeed: [Why Spotify Is Not Using Bayesian A/B Testing]() Original publisher: [Read original article]() Author: Spotify Engineering Published: 2026-09-08T13:18:44Z Content type: article Language: en Sources: [Spotify Engineering](), [Spotify Engineering Blog]() Topics: [Data Science](), [Tooling]() Tags: [a-b-testing](), [data](), [data-science](), [experimentation](), [inference](), [statistics](), [testing](), [tooling]() ### AI overview Spotify argues that Bayesian A/B testing can be oversimplified and that its claimed advantages require careful configuration. The article compares Bayesian and frequentist approaches and explains why Spotify does not currently need Bayesian inference alongside its frequentist tooling. ### Source excerpt Clearing the confusion about what Bayesian A/B testing is. The post Why Spotify Is Not Using Bayesian A/B Testing appeared first on Spotify Engineering. ## Introducing CUDA Rust: Two Tracks for Writing GPU Kernels DevFeed: [Introducing CUDA Rust: Two Tracks for Writing GPU Kernels]() Original publisher: [Read original article]() Author: Elizabeth Goodman Published: 2026-09-08T12:00:00Z Content type: article Language: en Sources: [NVIDIA Developer](), [NVIDIA Technical Blog]() Topics: [GPU](), [Code generation](), [toolchain](), [C++]() Tags: [c-plus-plus](), [cuda](), [cuda-tile](), [data-science](), [developer-tools-techniques](), [dynamo](), [gpu](), [kernels](), [nemo-retriever](), [open-source](), [programming-languages-compilers](), [rust](), [toolchains]() ### AI overview NVIDIA introduces CUDA Rust for writing GPU kernels natively in Rust and compiling them to PTX. The article explains the SIMT and Tile programming tracks, recommends Tile as the default, and describes cuda-oxide's Rust-based code-generation path. ### Source excerpt In September 2026, NVIDIA announced it is leaning into native GPU programming in Rust. CUDA C++ and CUDA Python are mature, enterprise-grade toolchains, and... ## Progressive Authentication for Production-Ready RAG Architectures DevFeed: [Progressive Authentication for Production-Ready RAG Architectures]() Original publisher: [Read original article]() Author: Aishwarya Murali Padikkal Published: 2026-09-08T07:16:00Z Content type: article Language: en Sources: [Data Science at Microsoft]() Topics: [Authentication](), [Retrieval Augmented Generation (RAG)](), [sensitive data](), [User Experience](), [Security]() Tags: [ai](), [ai-agent](), [authentication](), [data-science](), [privacy](), [rag](), [retrieval](), [sensitive-data](), [user-experience]() ### AI overview The article presents progressive authentication for RAG systems: answer public questions without sign-in, then authenticate within the conversation when a request requires access to sensitive personal data. It argues that static, upfront authentication does not fit conversations that shift between public and private contexts. ### Source excerpt Designing a production-ready RAG architecture where authentication becomes part of the conversation, not a prerequisite for it. By Aishwarya Murali Padikkal & Priyank Solanki From understanding to retrieval, grounding, and response, authentication enters only when the conversation requires access to sensitive data- illustrating the core idea of progressive authentication. Illustration generated using ChatGPT and refined by the authors. Two questions, one help line Imagine two taxpayers opening the same AI-powered tax assistant. The first asks, "When is the filing deadline this year?" The second asks, "What was my refund last year?" Both questions are valid. Both deserve fast, helpful answers. Yet they represent fundamentally different types of requests. Figure 1. The same assistant, two different journeys. Requests are routed according to the type of information being accessed. Illustration generated by AI (ChatGPT) and refined by the authors. The answer to the first lives in a public handbook. Anyone should be able to access it instantly- no sign-in, no friction. The answer to the second is within an individual's private tax records. Before revealing a single detail, the assistant must be certain it is interacting with the right person. Making public information instantly accessible while ensuring private information is revealed only to the right person, all within the same conversation, is the challenge we set out to solve. It is also where conventional RAG systems quietly fall apart. Traditional RAG architectures assume a much simpler interaction model: a user asks a question, the system retrieves relevant documents, and the model generates an answer. Authentication, if it exists at all, is treated as a one-time checkpoint at the start of the interaction. Real conversations are far less predictable. Users naturally shift contexts, starting with a public question, moving to a personal one, and then asking follow-up questions that rely on both. The conversation evo ## How to build a secure-by-default AI coding agent DevFeed: [How to build a secure-by-default AI coding agent]() Original publisher: [Read original article]() Author: Phoebe Sajor Published: 2026-09-04T07:40:00Z Content type: tutorial Language: en Sources: [Stack Overflow Blog]() Topics: [AI Bots]() Tags: [acquisitions](), [agent](), [ai](), [ai-coding](), [ai-infrastructure](), [coding](), [data](), [data-science](), [engineering](), [enterprise](), [how-to](), [infrastructure](), [learning](), [machine-learning](), [platform](), [podcast](), [products](), [python](), [se-stackoverflow](), [se-tech](), [security](), [software](), [supply-chain]() ### AI overview An interview about building a secure-by-default AI coding agent, the limits of prompts as security guardrails, and securing the AI software supply chain. ### Source excerpt Ryan chats with Greg Jennings, VP of Engineering for AI Products at Anaconda, about what it takes to build a secure-by-default AI coding agent, why prompts shouldn't be treated as strict security guardrails, and how Anaconda is using strategic acquisitions to secure the AI software supply chain. ## 9 GPU Concepts Every AI Engineer Should Know DevFeed: [9 GPU Concepts Every AI Engineer Should Know]() Original publisher: [Read original article]() Author: Dr. Ashish Bamania Published: 2026-08-31T15:19:52Z Content type: tutorial Language: en Sources: [Dr. Ashish Bamania]() Topics: [GPU](), [PyTorch](), [Large Language Model](), [Artificial Intelligence]() Tags: [ai-engineer](), [artificial-intelligence](), [coding](), [compute](), [cpu](), [data-science](), [gpu](), [guide](), [internals](), [llms](), [machine-learning](), [nvidia](), [programming](), [pytorch](), [technology]() ### AI overview A practical guide to nine GPU concepts relevant to AI engineers. It introduces GPU computation hardware, memory, program execution, and multi-GPU connections, including streaming multiprocessors in NVIDIA GPUs and their role in parallel computation. ### Source excerpt A simple and practical guide to GPU internals that AI engineers actually need. Continue reading on Level Up Coding " ## The Real Python Podcast - Episode #309: Exploring Complex Systems & Maintainable Data Science Pipelines DevFeed: [The Real Python Podcast - Episode #309: Exploring Complex Systems & Maintainable Data Science Pipelines]() Original publisher: [Read original article]() Author: Real Python Published: 2026-08-28T12:00:00Z Content type: article Language: en Sources: [Real Python]() Topics: [Complex Systems](), [Data Science](), [Python](), [systems](), [configuration](), [ci](), [YAML](), [releases](), [Django](), [VS Code Extension]() Tags: [code](), [complex-systems](), [configuration](), [data-science](), [django](), [podcast](), [python](), [release](), [systems](), [vs-code](), [workflow]() ### AI overview This Real Python podcast episode explores complex versus complicated coding problems and practical patterns for designing maintainable systems. It also covers repeatable data science pipelines, DataFrame validation with Pointblank, configuration-driven modular workflows, Python releases, PEPs, Django's release cycle, and several Python community projects. ### Source excerpt What are the key characteristics of complex systems, and what are practical patterns for tackling complex coding problems? Christopher Trudeau is back on the show this week with another batch of PyCoder's Weekly articles and projects. ## LLMs Are Not (Consistently) Bayesian: Quantifying Internal (In)consistencies of LLMs' Probabilistic Beliefs DevFeed: [LLMs Are Not (Consistently) Bayesian: Quantifying Internal (In)consistencies of LLMs' Probabilistic Beliefs]() Original publisher: [Read original article]() Published: 2026-08-28T00:00:00Z Content type: article Language: en Sources: [Apple Machine Learning Research]() Topics: [Large Language Model](), [Uncertainty quantification LLMs](), [AI Inference](), [Data Science]() Tags: [ai](), [data-science](), [diagnostics](), [llms](), [research]() ### AI overview This research examines whether large language models update probabilistic beliefs in accordance with Bayes' rule. It introduces the information processing gap to quantify deviations from Bayesian updates, compares evidence-integration approaches, and finds that heuristic, non-Bayesian updates can outperform exact Bayesian updates on downstream tasks. ### Source excerpt Modern AI systems are being deployed in complex domains such as medicine, science, and law, where there is often not a single correct answer given the observed evidence. Such systems must be able to represent and update uncertain beliefs about the world as new evidence arrives to make rational decisions. We introduce the novel technique of studying LLMs as information processing rules and utilize the information processing gap--the deviation from Bayes updates--to study the internal (in)consistencies of how LLMs update their probabilistic beliefs from evidence. Our extensive experiments evaluate... ## Restore LLM Inference Capacity in Seconds with Shadow Engine Recovery in NVIDIA Dynamo DevFeed: [Restore LLM Inference Capacity in Seconds with Shadow Engine Recovery in NVIDIA Dynamo]() Original publisher: [Read original article]() Author: Michelle Horton Published: 2026-08-25T20:57:54Z Content type: article Language: en Sources: [NVIDIA Developer](), [NVIDIA Technical Blog]() Topics: [Large Language Model](), [Inference](), [model-deployment](), [Nvidia](), [Optimization](), [CUDA](), [Processes](), [GPU](), [NCCL](), [sglang]() Tags: [agentic-ai-generative-ai](), [cuda](), [cuda-x](), [data-science](), [deployment](), [developer-tools-techniques](), [gpu](), [inference](), [llm](), [models](), [nccl](), [nvidia](), [optimization](), [process]() ### AI overview This article explains NVIDIA Dynamo's shadow engine recovery for LLM inference. A fully initialized shadow engine shares weights with the active engine and takes over after a process failure, reducing recovery from a 283-second cold restart to 7.3 seconds in the described two-worker deployment. ### Source excerpt When an LLM engine process fails, the standard recovery path involves a cold restart. This requires loading weights into HBM from storage, compiling kernels,... ## Using Data Contracts to Coordinate Data Evolution at Enterprise Scale DevFeed: [Using Data Contracts to Coordinate Data Evolution at Enterprise Scale]() Original publisher: [Read original article]() Author: Keerthipriyan Published: 2026-08-25T20:22:17Z Content type: article Language: en Sources: [Walmart Global Tech]() Topics: [data-platforms](), [Data Quality](), [DataOps](), [systems]() Tags: [architecture](), [data](), [data-engineering](), [data-governance](), [data-platforms](), [data-quality](), [data-science](), [enterprise](), [organizational](), [pipeline](), [schema](), [software-engineering](), [teams](), [technology]() ### AI overview The article explains how data contracts help large enterprises coordinate changes across independently evolving data teams and downstream consumers. It argues that schema validation alone cannot identify ownership, downstream impact, or migration responsibilities, and presents data contracts as machine-enforceable coordination agreements. ### Source excerpt Coauthored by Satyajeet Coordinating Data Evolution at Enterprise Scale When you operate data platforms on a global enterprise scale, hundreds of engineering teams ship improvements every week, each moving independently to deliver value at the pace of the business demands. This velocity is a competitive advantage. The challenge: How do you enable hundreds of teams to evolve their data products independently while maintaining reliability for thousands of downstream consumers? Traditional coordination methods (messages, wiki updates, shared spreadsheets) work at small scale but break at Walmart scale. A source team ships an enhancement, perfectly valid within their domain, but that change ripples through fifteen downstream pipelines owned by different teams with different release schedules. Without a formal coordination mechanism, you discover the impact after it reaches production. The gap isn't technical debt or fragile systems. It's the absence of machine-enforceable agreements that scale with organizational complexity. Data contracts solve this: enabling teams to move fast independently while maintaining coordinated reliability across organizational boundaries. Here's the architecture we built. Why Schema Validation Alone Isn't Enough When data quality issues surface in production, the first instinct is often added to more schema validation. If a field is missing or has the wrong type, the pipeline catches it. This works for many data quality problems, but not all of them. Consider a scenario where a source team enhances their data model by restructuring field names to support new business capabilities. The schema still validates perfectly: every field exists; every type is correct; the data is well formed. But downstream consumers who depend on the original field names now receive empty results. Schema validation checks whether data has the right shape. It tells you that a field is missing. It does not tell you who owns that field, which downstream teams will bre ## CUDA Python 1.0: Stable APIs, One Foundation, Full Platform Access DevFeed: [CUDA Python 1.0: Stable APIs, One Foundation, Full Platform Access]() Original publisher: [Read original article]() Author: Elizabeth Goodman Published: 2026-08-25T15:00:00Z Content type: article Language: en Sources: [NVIDIA Developer](), [NVIDIA Technical Blog]() Topics: [CUDA](), [Python](), [GPU](), [C++](), [toolchain](), [PyTorch](), [RAPIDS](), [releases]() Tags: [c-plus-plus](), [cuda](), [data-science](), [developer-tools-techniques](), [gpu](), [numba](), [python](), [pytorch](), [rapids](), [release](), [toolchain]() ### AI overview CUDA Python 1.0 makes the CUDA platform accessible from Python through stable, versioned components, including runtime access, low-level CUDA C API bindings, parallel algorithms, component discovery, and NVIDIA math libraries. The release establishes a shared foundation for Python GPU libraries and introduces semantic-versioning commitments for public APIs. ### Source excerpt For years, a Python developer who needed a GPU had two realistic choices: Learn NVIDIA CUDA C++ well enough to write an extension, set up a build toolchain, and... ## How does Claude watermark text? DevFeed: [How does Claude watermark text?]() Original publisher: [Read original article]() Author: Dr. Ashish Bamania Published: 2026-08-24T15:20:14Z Content type: tutorial Language: en Sources: [Dr. Ashish Bamania]() Topics: [Claude](), [watermarking](), [anthropic](), [Large Language Model](), [tokenization]() Tags: [anthropic](), [artificial-intelligence](), [claude](), [data-science](), [llms](), [programming](), [software-engineering](), [technology](), [tokens](), [watermarking]() ### AI overview A visual tutorial explains how Anthropic plans to watermark and detect text generated by Claude models, and introduces token-by-token text generation, token IDs, embeddings, logits, and probabilities. ### Source excerpt A visual lesson to clearly understand how Anthropic watermarks and detects Claude-generated text. Continue reading on Level Up Coding " ## GPU-Accelerated Clustering for Financial Instruments at Scale DevFeed: [GPU-Accelerated Clustering for Financial Instruments at Scale]() Original publisher: [Read original article]() Author: Elizabeth Goodman Published: 2026-08-21T16:21:04Z Content type: article Language: en Sources: [NVIDIA Developer](), [NVIDIA Technical Blog]() Topics: [GPU](), [Algorithm](), [Matrix]() Tags: [batch](), [communication](), [data-analytics-processing](), [data-science](), [financial-services](), [gb200](), [gpu](), [memory](), [nemo](), [nemotron](), [nvidia](), [post](), [scale](), [simulation-modeling-design](), [workflow]() ### AI overview A GPU-accelerated workflow uses rolling correlation and tail-dependence matrices to cluster financial instruments for portfolio construction, risk aggregation, statistical arbitrage, and trade surveillance. Its adaptive SymNMF-based solver supports soft factor loadings and hard cluster labels, while memory-efficient and distributed implementations scale from single GPUs to one million instruments across multiple nodes. ### Source excerpt Use AdaptGrow, a GPU-accelerated matrix factorization algorithm, to turn rolling correlation and tail-dependence matrices into hard clusters, soft factor... ## Building Federated Multimodal AI Workflows with NVIDIA FLARE DevFeed: [Building Federated Multimodal AI Workflows with NVIDIA FLARE]() Original publisher: [Read original article]() Author: Tanya Lenz Published: 2026-08-19T17:50:47Z Content type: article Language: en Sources: [NVIDIA Developer](), [NVIDIA Technical Blog]() Topics: [vlm](), [SDKs]() Tags: [agentic-ai-generative-ai](), [ai](), [data-science](), [edge-computing](), [featured](), [federated-learning](), [multimodal](), [multimodal-ai](), [nvidia](), [python](), [streaming](), [training](), [training-ai-models](), [updates](), [vlms](), [workflows]() ### AI overview The article explains how NVIDIA FLARE supports federated training for multimodal and vision-language models when data remains distributed across sites. It focuses on deciding which model state to exchange and on efficiently transferring and aggregating large updates through externalization, tensor streaming, and disk-backed aggregation. ### Source excerpt Modern vision-language models (VLMs) can support tasks such as visual question answering, captioning, and image-text reasoning. In practice, however, the data... ## How we knew COVID was over (and what our models had to unlearn) DevFeed: [How we knew COVID was over (and what our models had to unlearn)]() Original publisher: [Read original article]() Author: Harrison Katz Published: 2026-08-19T17:01:03Z Content type: article Language: en Sources: [The Airbnb Tech Blog - Medium]() Topics: [Data Science](), [data](), [datasets](), [Process]() Tags: [company](), [data](), [data-modeling](), [data-science](), [machine-learning](), [models](), [process](), [technology]() ### AI overview An Airbnb forecasting data science team explains how it responds when production forecasts drift, distinguishing between refitting a model with newer data, respecifying its structure, and holding it unchanged. The article emphasizes diagnosing the source of persistent bias and managing the risks of model updates that influence company decisions. ### Source excerpt When we retrain, when we rebuild, and when we leave a model alone. By: Harrison Katz A forecast that carries weight The Forecasting Data Science team at Airbnb produces many of the forecasts the rest of the company plans around: demand, bookings, cancellations, and a range of finer cuts by market and segment, refreshed continuously across thousands of markets. The targets differ, and the models differ, but they have one thing in common: Other teams build on top of them. This means a forecast that is casually wrong is not a clean miss, as it might be in an academic setting. That's because a small bias does not stay small once a lot of decisions are riding on it. So when one of those forecasts starts to drift, what to do about it is not really a methods question. It is a risk question, and an easy one to get wrong, which we have from time to time. One of these forecasts had been missing, compared to what actually happened after the forecast was released, in the same direction for a couple of quarters. This bias persisted after several routine refreshes. The usual solution would be to fully retrain the model: pull in the recent data, refit the model again, and ship. But we wanted to understand the source of the bias, rather than simply hoping an update would eliminate it. If you're interested in other posts on this topic, you can learn more about how COVID impacted Airbnb's financial models or how we dealt with disruption to our models during the pandemic. This post is about the discipline that came out of both: how we now decide whether a struggling forecast needs new data, a new model, or no changes at all. One word, three decisions The easy mistake is treating the choice to "retrain" a model as a single action. It is three separate actions -- refitting, respecifying, or holding -- and none of them is particularly similar to the others. Refitting is the cheaper option. Same model, same structure, same features, updated with newer data. This is what most people mean by ## Run Massive-Scale UMAP in Minutes Using Multiple GPUs--Without Losing Accuracy DevFeed: [Run Massive-Scale UMAP in Minutes Using Multiple GPUs--Without Losing Accuracy]() Original publisher: [Read original article]() Author: Tanya Lenz Published: 2026-08-18T16:48:08Z Content type: article Language: en Sources: [NVIDIA Developer](), [NVIDIA Technical Blog]() Topics: [Data analysis](), [GPU](), [Graphs](), [datasets](), [RAPIDS](), [Nvidia]() Tags: [accelerate](), [analysis](), [cuda-x](), [data-analytics-processing](), [data-center-cloud](), [data-science](), [datasets](), [feature](), [featured](), [gpu](), [graphs](), [multi-gpu](), [nvidia](), [post](), [scale](), [simulation-modeling-design](), [training](), [vector](), [visualization]() ### AI overview This article explains how multi-GPU UMAP scales dimensionality reduction to datasets containing tens to hundreds of millions of vectors. A feature in NVIDIA cuML and cuVS 25.06 distributes all-neighbors kNN graph construction across multiple GPUs, enabling workloads of several hundred gigabytes to run in minutes while preserving nearest-neighbor relationships and accuracy. ### Source excerpt Uniform Manifold Approximation and Projection (UMAP) is a dimensionality reduction technique widely used for visualization and feature extraction. Applications... ## Beyond Relevance: Building a Quality-Aware Retrieval Layer for RAG DevFeed: [Beyond Relevance: Building a Quality-Aware Retrieval Layer for RAG]() Original publisher: [Read original article]() Author: Shay Ben-Elazar Published: 2026-08-18T07:16:01Z Content type: article Language: en Sources: [Data Science at Microsoft]() Topics: [Retrieval Augmented Generation (RAG)](), [Large Language Model](), [Artificial Intelligence]() Tags: [ai](), [data-science](), [education](), [machine-learning](), [microsoft](), [quality](), [rag](), [retrieval]() ### AI overview The article examines how a metadata-aware quality layer could improve educational RAG systems by ranking or filtering retrieved sources for reliability, clarity, evidence, and suitability for teaching. It argues that relevance alone does not ensure educational quality and that filtering weak sources can also reduce context noise, token use, and inference costs. ### Source excerpt How a metadata-aware quality layer can help educational RAG systems retrieve sources that are relevant, reliable, and ready for teaching Shay Ben-Elazar, Principal Applied Data Science Manager, Microsoft Rob Mauceri, Distinguished Engineer, Microsoft A student asks an AI tutor a science question: "Why do flu vaccines need to be updated over time?" A retrieval-augmented generation (RAG) system searches its index, retrieves passages about influenza viruses, vaccine development, immunity, and seasonal outbreaks, then generates an answer grounded in those sources [1]. At first glance, the system appears to have done its job. The passages are relevant, and the answer cites supporting information. But a deeper question remains: were the retrieved sources actually good sources to learn from? Some passages may be outdated, overly technical, weakly supported, or poorly organized for a student audience. Others may provide clearer explanations, stronger evidence, and a more coherent path to understanding. Both may be relevant, yet only one is likely to help a learner build accurate understanding [2]. This distinction matters because relevance alone is not the same as educational quality. Retrieved passages can match the topic while lacking clear definitions, supporting evidence, or an easy-to-follow explanation. In a student-facing product, those differences shape not only what the student learns, but also how much they trust the system. There is also a practical systems reason to care. Web-scale RAG applications operate within limited context windows and real compute budgets. Filtering weak sources earlier, or downranking them before generation, reduces noise, preserves tokens for stronger evidence, and lowers inference costs, compounding across products used by millions of learners [3]. RAG has become one of the most practical ways to make large language model applications more grounded. It can make answers more current, more domain-specific, and easier to connect back to so ## Confluent Cloud for Apache Flink: Engine for Mission-Critical, Real-Time Operational Systems and dbt/SQL-Native Home for Data Science and AI DevFeed: [Confluent Cloud for Apache Flink: Engine for Mission-Critical, Real-Time Operational Systems and dbt/SQL-Native Home for Data Science and AI]() Original publisher: [Read original article]() Author: Yashwanth Dasari Published: 2026-08-18T02:20:00Z Content type: release Language: en Sources: [Confluent: Data in motion]() Topics: [Confluent Cloud](), [apache-flink](), [stream-processing](), [Kafka](), [AI, ML & Data Engineering](), [Cloud](), [SDKs]() Tags: [ai](), [announcement](), [apache-flink](), [api](), [batch](), [confluent-cloud](), [data-science](), [developers](), [kafka](), [real-time]() ### AI overview Confluent announces an evolution of Confluent Cloud for Apache Flink that unifies mission-critical real-time operations with analytics and AI workflows. The release adds a serverless, co-designed Kafka and Flink platform, including the generally available Flink Table API in Java for code-first development. ### Source excerpt Flink now acts as a robust engine for developers through Table API, UDFs, and PTFs while offering a SQL-native, dbt-integrated platform for data science and AI teams. ## When Can LLMs Replace Humans in A/B Tests? DevFeed: [When Can LLMs Replace Humans in A/B Tests?]() Original publisher: [Read original article]() Author: Spotify Engineering Published: 2026-08-13T18:57:22Z Content type: article Language: en Sources: [Spotify Engineering](), [Spotify Engineering Blog]() Topics: [Large Language Model](), [A/B Testing](), [datasets](), [data](), [experiments]() Tags: [a-b-testing](), [analysis](), [data](), [data-science](), [experiment](), [gpt](), [llm](), [llms](), [model](), [testing]() ### AI overview The article examines whether large language model predictions can replace human outcomes in A/B tests. Using the Upworthy dataset, it finds that calibrated predictions can recover treatment effects under specific assumptions, while raw predictions recovered only 39% of the observed human effect. ### Source excerpt TL;DR: LLM predictions can stand in for human outcomes in A/B tests, but only by assumption, not by design.... The post When Can LLMs Replace Humans in A/B Tests? appeared first on Spotify Engineering. ## Whatnot at Snowflake Summit 2026 DevFeed: [Whatnot at Snowflake Summit 2026]() Original publisher: [Read original article]() Author: Whatnot Engineering Published: 2026-08-12T17:54:54Z Content type: article Language: en Sources: [Whatnot Engineering]() Topics: [data](), [datasets](), [Artificial Intelligence](), [developer velocity]() Tags: [ai](), [ai-agent](), [analytics](), [data](), [data-science](), [datasets](), [developer-velocity](), [developers](), [machine-learning]() ### AI overview Whatnot's Data Platform team describes the data platform behind its live marketplace and its use of Snowflake Cortex AI and Semantic Views. The approach grounds AI in governed business definitions and trusted metadata so developers can find datasets and metrics and investigate issues using natural-language questions. ### Source excerpt Earlier this summer, several members of the Whatnot Data Platform team joined Snowflake on stage at Snowflake Summit to share how we're building the data platform behind Whatnot's live marketplace. This post covers what we talked about, why we built it, and future direction. Caption: The Whatnot team were everywhere at the Summit! Snowflake recently published a recap of our sessions, which you can read here: Observability at Scale: Whatnot at Snowflake Summit. Data at Whatnot Every live auction, bid, chat message, purchase, moderation action and recommendation produces data. That translates into billions of events each day powering recommendations, seller analytics, and customer support. For us, data isn't just for internal analysis. We use it to power the product while millions of people are buying and selling in real time. That creates an unique set of engineering challenges. As Whatnot has grown, we've had to evolve our platform to support thousands of datasets, hundreds of developers, and an increasingly diverse set of workloads while maintaining reliability, governance, and developer velocity. At Summit, we talked about the investments we have made to get engineers to vetted data faster Accelerating Builders with AI The Whatnot data platform is a living system. Schemas evolve, pipelines are deployed continuously, business definitions change, and new data products are published every day. Keeping up with that pace of change has become one of the biggest challenges for developers. In our first session, we shared how we're using Snowflake Cortex AI and Semantic Views to make that knowledge more accessible. By grounding AI in governed business definitions and trusted metadata (see this earlier post for more details), builders can ask natural language questions about their data, quickly identify the right datasets and metrics, and investigate issues without manually searching documentation or relying on the Data team. Whatnot's full-time employees now use AI weekly, ## Calibrating LLM-Based Population Estimates with Human Validation DevFeed: [Calibrating LLM-Based Population Estimates with Human Validation]() Original publisher: [Read original article]() Author: Hiroshi Urata Published: 2026-08-12T00:29:34Z Content type: article Language: en Sources: [Indeed]() Topics: [Large Language Model](), [human feedback](), [data](), [LLM evaluation / benchmarking]() Tags: [classification](), [data](), [data-science](), [dataset](), [evaluation](), [false-negative](), [false-positive](), [llm](), [machine-learning](), [measurement](), [validation]() ### AI overview The article explains how human validation can calibrate LLM-based population estimates. It distinguishes an LLM's apparent positive rate from true prevalence, uses human-validated samples to estimate sensitivity and specificity, and applies those error estimates to correct population-level measurements and quantify uncertainty. ### Source excerpt Key Idea Human validation is not only for evaluating an LLM. It can also calibrate how the LLM is used as a scalable measurement instrument for population estimation. An LLM can classify thousands of records at low cost, but the proportion it classifies as positive is not necessarily the true proportion in the population. By [...] ## Measuring time savings from Figma Make DevFeed: [Measuring time savings from Figma Make]() Original publisher: [Read original article]() Author: Remy Stewart Published: 2026-08-11T19:59:00Z Content type: article Language: en Sources: [Figma Blog]() Topics: [Figma](), [Data Science](), [Artificial Intelligence](), [A/B Testing]() Tags: [a-b-testing](), [ai](), [ai-tools](), [data-science](), [figma](), [productivity](), [research](), [testing]() ### AI overview Figma's Data Science team describes a randomized controlled trial measuring how Figma Make affects users' design work. Across 100 participants, design work was 20% faster and 16% easier overall, while PMs saw tasks become 23% faster and 37% easier. The article explains why confounders make AI time savings difficult to measure and discusses the limitations of online A/B testing and causal inference using log data. ### Source excerpt The Figma Data Science team assumed that AI saves users time--but quantifying it required a new approach to research design. [Next page]()