# Airbnb Engineering & Data Science

Airbnb Tech Blog

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

## De-identifying text at scale: How Airbnb protects personal data in AI training

DevFeed: [De-identifying text at scale: How Airbnb protects personal data in AI training](<https://devfeed.tech/articles/de-identifying-text-at-scale-how-airbnb-protects-personal-data-in-ai-training-141090.md>)

Original publisher: [Read original article](<https://airbnb.tech/infrastructure/de-identifying-text-at-scale-how-airbnb-protects-personal-data-in-ai-training/>)

Author: natalijafijacko

Published: 2026-11-18T16:33:01Z

Content type: article

Language: en

Sources: [Airbnb Engineering & Data Science](<https://devfeed.tech/sources/airbnb-engineering-data-science.md>)

Topics: [managing large-scale text data in AI](<https://devfeed.tech/topics/managing-large-scale-text-data-in-ai.md>), [data loss prevention](<https://devfeed.tech/topics/data-loss-prevention.md>), [Data Quality](<https://devfeed.tech/topics/data-quality.md>), [ner](<https://devfeed.tech/topics/ner.md>), [ML model versioning](<https://devfeed.tech/topics/ml-model-versioning.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [ai-training](<https://devfeed.tech/tags/ai-training.md>), [airbnb](<https://devfeed.tech/tags/airbnb.md>), [data](<https://devfeed.tech/tags/data.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [infrastructure](<https://devfeed.tech/tags/infrastructure.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [personal](<https://devfeed.tech/tags/personal.md>), [personal-data](<https://devfeed.tech/tags/personal-data.md>), [scale](<https://devfeed.tech/tags/scale.md>), [technology](<https://devfeed.tech/tags/technology.md>), [text](<https://devfeed.tech/tags/text.md>), [training](<https://devfeed.tech/tags/training.md>)

### AI overview

Airbnb describes an in-process text de-identification library built on Presidio to remove personal data before ML training. Teams provide data schemas; Airbnb supplies recognizers, operators, language handling and runtime configuration, while dataset tags and recorded versions/configurations support production gating and traceability. Detection and replacement are separate stages: Airbnb adds phrase-aware context for identifiers such as listing IDs and gift cards, and uses GLiNER for person and street-address detection where spaCy's broad location tags were insufficient. Replacement usually uses realistic Faker values; run-specific secrets make outputs consistent within a run but different across runs, without storing a real-to-fake mapping. Other data receives tailored handling, including salted hashes, coordinate truncation and format-preserving encryption for internal IDs. Detection thresholds and replacement methods vary by use case. Quality is measured by entity and language for recall, precision and throughput, with recall favored for privacy; benchmark comparisons gate model and library changes against regressions. The account covers text deployment and says work on other-媒体

### Source excerpt

Here's how Airbnb detects and replaces personal data in messages, reviews, and support chats before it reaches a model.

## Skipper, Airbnb's embedded workflow engine, is now open source

DevFeed: [Skipper, Airbnb's embedded workflow engine, is now open source](<https://devfeed.tech/articles/skipper-airbnb-s-embedded-workflow-engine-is-now-open-source-129354.md>)

Original publisher: [Read original article](<https://airbnb.tech/open-source/skipper-airbnbs-embedded-workflow-engine-is-now-open-source/>)

Author: natalijafijacko

Published: 2026-11-10T18:08:34Z

Content type: article

Language: en

Sources: [Airbnb Engineering & Data Science](<https://devfeed.tech/sources/airbnb-engineering-data-science.md>)

Topics: [Orchestration](<https://devfeed.tech/topics/orchestration.md>), [workflow automation](<https://devfeed.tech/topics/workflow-automation.md>), [Open Source](<https://devfeed.tech/topics/open-source.md>)

Tags: [actions](<https://devfeed.tech/tags/actions.md>), [adapters](<https://devfeed.tech/tags/adapters.md>), [airbnb](<https://devfeed.tech/tags/airbnb.md>), [durable-workflows](<https://devfeed.tech/tags/durable-workflows.md>), [embedded](<https://devfeed.tech/tags/embedded.md>), [engine](<https://devfeed.tech/tags/engine.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [open](<https://devfeed.tech/tags/open.md>), [open-source](<https://devfeed.tech/tags/open-source.md>), [orchestration](<https://devfeed.tech/tags/orchestration.md>), [service](<https://devfeed.tech/tags/service.md>), [skipper](<https://devfeed.tech/tags/skipper.md>), [software-architecture](<https://devfeed.tech/tags/software-architecture.md>), [source](<https://devfeed.tech/tags/source.md>), [technology](<https://devfeed.tech/tags/technology.md>), [workflow](<https://devfeed.tech/tags/workflow.md>), [workflow-engine](<https://devfeed.tech/tags/workflow-engine.md>), [workflows](<https://devfeed.tech/tags/workflows.md>)

### AI overview

Skipper is an Apache 2.0 Java/Kotlin workflow engine that runs inside a service, persisting workflow state and scheduled work in a shared database rather than requiring a separate orchestration cluster. Replicas run its runtime and scheduler, which coordinate through leases and can recover work after a replica fails; the service team owns execution, database capacity, and recovery operations. Workflows are replayed from the top using saved action checkpoints, so orchestration must be deterministic and actions must make side effects idempotent. Batched checkpoints reduce database writes but can cause completed actions to run again after a crash; even immediate checkpoints cannot make downstream effects atomic. Retries, signals, waits, compensation, and workflow tests are supported, while permanent failures or exhausted recovery attempts may need handling or manual intervention. The release includes MySQL and SQLite adapters, an admin UI, test utilities, and a Kotlin state-machine DSL; additional storage plugins are described as future work.

### Source excerpt

Durable workflows, running inside your service with no separate orchestration cluster.

## Frontier quality at 1% of the cost: a production benchmark of LLM classifiers

DevFeed: [Frontier quality at 1% of the cost: a production benchmark of LLM classifiers](<https://devfeed.tech/articles/frontier-quality-at-1-of-the-cost-a-production-benchmark-of-llm-classifiers-116278.md>)

Original publisher: [Read original article](<https://airbnb.tech/ai-ml/frontier-quality-at-1-of-the-cost-a-production-benchmark-of-llm-classifiers/>)

Author: natalijafijacko

Published: 2026-11-03T22:31:31Z

Content type: news

Language: en

Sources: [Airbnb Engineering & Data Science](<https://devfeed.tech/sources/airbnb-engineering-data-science.md>)

Topics: [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>), [model-deployment](<https://devfeed.tech/topics/model-deployment.md>), [Benchmark](<https://devfeed.tech/topics/benchmark.md>), [Large Language Model](<https://devfeed.tech/topics/llm.md>), [Frontier AI](<https://devfeed.tech/topics/frontier-ai.md>), [cloud cost management](<https://devfeed.tech/topics/cloud-cost-management.md>)

Tags: [ai-ml](<https://devfeed.tech/tags/ai-ml.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [classifiers](<https://devfeed.tech/tags/classifiers.md>), [cost](<https://devfeed.tech/tags/cost.md>), [data](<https://devfeed.tech/tags/data.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [frontier](<https://devfeed.tech/tags/frontier.md>), [infrastructure](<https://devfeed.tech/tags/infrastructure.md>), [llm](<https://devfeed.tech/tags/llm.md>), [production](<https://devfeed.tech/tags/production.md>), [quality](<https://devfeed.tech/tags/quality.md>), [technology](<https://devfeed.tech/tags/technology.md>)

### AI overview

On Airbnb's pricing and payment transparency screening task, fine-tuning Gemma4-E2B raised max F1 from 0.563 to 0.902, with 35 ms median latency and about 85 ms p99, matching the strongest proprietary models' quality at much lower serving cost when traffic is high. The benchmark used 10,000 weighted held-out examples; open-weight models were self-hosted on one NVIDIA RTX Pro 6000 GPU, while proprietary models were called through vendor APIs. Off-the-shelf open-weight models were fast but less accurate; proprietary generative models were accurate but typically much slower and returned fixed verdicts, limiting precision-recall control. TypeSafe Jev returned thresholdable probabilities and was the strongest off-the-shelf model under 100 ms. A detailed 2,700-word prompt modestly helped fast models but did not close their quality gap, while increasing latency and cost. Self-hosting was cheaper above roughly 15-30 RPM for some frontier APIs and around 30 RPM overall; the authors say frontier models may suit low-volume or frequently changing tasks. Cost estimates use public rates and favor hosted models through an assumed 71% cache hit rate.

### Source excerpt

We benchmarked a fine-tuned E2B model against frontier APIs on a real production classification task. It matched their quality in under 100ms, at about a hundredth of the cost.

## CASTLE: training natural language search before the first real query

DevFeed: [CASTLE: training natural language search before the first real query](<https://devfeed.tech/articles/castle-training-natural-language-search-before-the-first-real-query-86070.md>)

Original publisher: [Read original article](<https://airbnb.tech/ai-ml/castle-training-natural-language-search-before-the-first-real-query/>)

Author: natalijafijacko

Published: 2026-10-13T17:36:30Z

Content type: article

Language: en

Sources: [Airbnb Engineering & Data Science](<https://devfeed.tech/sources/airbnb-engineering-data-science.md>)

Topics: [text2sql](<https://devfeed.tech/topics/text2sql.md>), [Large Language Model](<https://devfeed.tech/topics/llm.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [cosine similarity search](<https://devfeed.tech/topics/cosine-similarity-search.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [ai-ml](<https://devfeed.tech/tags/ai-ml.md>), [annotation](<https://devfeed.tech/tags/annotation.md>), [data](<https://devfeed.tech/tags/data.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [relevance](<https://devfeed.tech/tags/relevance.md>), [search](<https://devfeed.tech/tags/search.md>), [synthetic](<https://devfeed.tech/tags/synthetic.md>), [technology](<https://devfeed.tech/tags/technology.md>), [training](<https://devfeed.tech/tags/training.md>), [training-data](<https://devfeed.tech/tags/training-data.md>)

### AI overview

Airbnb's CASTLE framework tackles the cold-start problem for natural-language search by jointly generating synthetic queries and relevance labels from de-identified booking sessions. Contrastive pairs of viewed and booked listings ground the labels in user behavior, while several prompting methods generate varied, realistic queries. The article reports better alignment with real query distributions than the compared baselines and 91-93% agreement from expert reviewers on sampled relevance assignments.

### Source excerpt

How we addressed the cold-start problem for Airbnb's natural language search by jointly generating synthetic training data -- queries and relevance labels -- from de-identified booking and search signals.

## Beyond synthetic testing: Capturing and replaying real database workloads at Airbnb

DevFeed: [Beyond synthetic testing: Capturing and replaying real database workloads at Airbnb](<https://devfeed.tech/articles/beyond-synthetic-testing-capturing-and-replaying-real-database-workloads-at-airbnb-65578.md>)

Original publisher: [Read original article](<https://airbnb.tech/infrastructure/beyond-synthetic-testing-capturing-and-replaying-real-database-workloads-at-airbnb/>)

Author: natalijafijacko

Published: 2026-10-06T17:23:49Z

Content type: article

Language: en

Sources: [Airbnb Engineering & Data Science](<https://devfeed.tech/sources/airbnb-engineering-data-science.md>)

Topics: [Performance Testing](<https://devfeed.tech/topics/performance-testing.md>), [Database](<https://devfeed.tech/topics/database.md>), [read-replica](<https://devfeed.tech/topics/read-replica.md>), [Transactions](<https://devfeed.tech/topics/transactions.md>)

Tags: [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [bin-packing](<https://devfeed.tech/tags/bin-packing.md>), [capacity-planning](<https://devfeed.tech/tags/capacity-planning.md>), [concurrency](<https://devfeed.tech/tags/concurrency.md>), [cost-optimization](<https://devfeed.tech/tags/cost-optimization.md>), [data](<https://devfeed.tech/tags/data.md>), [database](<https://devfeed.tech/tags/database.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [infrastructure](<https://devfeed.tech/tags/infrastructure.md>), [sql](<https://devfeed.tech/tags/sql.md>), [technology](<https://devfeed.tech/tags/technology.md>), [testing](<https://devfeed.tech/tags/testing.md>)

### AI overview

Airbnb describes a system that captures production MySQL-compatible database traffic through ProxySQL, processes it into per-cluster datasets, and replays it offline. The system supports load testing with real workloads, capacity planning, performance debugging, and compatibility checks for database upgrades and migrations. Its replay modes either send traffic to one target or compare results across two targets.

### Source excerpt

How we capture real production database traffic at Airbnb and replay it offline to load-test, plan capacity, and de-risk upgrades.

## Personalization without user identity

DevFeed: [Personalization without user identity](<https://devfeed.tech/articles/personalization-without-user-identity-62263.md>)

Original publisher: [Read original article](<https://airbnb.tech/infrastructure/personalization-without-user-identity/>)

Author: natalijafijacko

Published: 2026-09-29T21:32:38Z

Content type: article

Language: en

Sources: [Airbnb Engineering & Data Science](<https://devfeed.tech/sources/airbnb-engineering-data-science.md>)

Topics: [personalization](<https://devfeed.tech/topics/personalization.md>), [search-engines](<https://devfeed.tech/topics/search-engines.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [ai-ml](<https://devfeed.tech/tags/ai-ml.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [identifier](<https://devfeed.tech/tags/identifier.md>), [infrastructure](<https://devfeed.tech/tags/infrastructure.md>), [location](<https://devfeed.tech/tags/location.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [ml](<https://devfeed.tech/tags/ml.md>), [personalization](<https://devfeed.tech/tags/personalization.md>), [privacy](<https://devfeed.tech/tags/privacy.md>), [ranking](<https://devfeed.tech/tags/ranking.md>), [recommendations](<https://devfeed.tech/tags/recommendations.md>), [technology](<https://devfeed.tech/tags/technology.md>)

### AI overview

Airbnb describes Proximity Features, which group activity from nearby users into geographic buckets to personalize search ranking, destination recommendations, and marketing pages for visitors with little or no individual history. The approach uses signals available when an IP address can be geo-located and is designed to avoid persistent user identifiers, amid privacy regulations and reduced access to third-party cookies.

### Source excerpt

How Airbnb uses proximity signals to personalize without relying on individual user history.

## The guest journey, updated in real time: extending Airbnb's sequence recommender with Chronon

DevFeed: [The guest journey, updated in real time: extending Airbnb's sequence recommender with Chronon](<https://devfeed.tech/articles/the-guest-journey-updated-in-real-time-extending-airbnb-s-sequence-recommender-with-chronon-59053.md>)

Original publisher: [Read original article](<https://airbnb.tech/ai-ml/the-guest-journey-updated-in-real-time-extending-airbnbs-sequence-recommender-with-chronon/>)

Author: laurenmackevich

Published: 2026-09-23T20:21:27Z

Content type: article

Language: en

Sources: [Airbnb Engineering & Data Science](<https://devfeed.tech/sources/airbnb-engineering-data-science.md>)

Topics: [real-time](<https://devfeed.tech/topics/real-time.md>), [Latency](<https://devfeed.tech/topics/latency.md>), [Embeddings](<https://devfeed.tech/topics/embeddings.md>), [Inference](<https://devfeed.tech/topics/inference.md>), [Transformer](<https://devfeed.tech/topics/transformer.md>), [Open Source](<https://devfeed.tech/topics/open-source.md>), [data](<https://devfeed.tech/topics/data.md>)

Tags: [ai-ml](<https://devfeed.tech/tags/ai-ml.md>), [batch](<https://devfeed.tech/tags/batch.md>), [capabilities](<https://devfeed.tech/tags/capabilities.md>), [data](<https://devfeed.tech/tags/data.md>), [embeddings](<https://devfeed.tech/tags/embeddings.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [inference](<https://devfeed.tech/tags/inference.md>), [infrastructure](<https://devfeed.tech/tags/infrastructure.md>), [latency](<https://devfeed.tech/tags/latency.md>), [low-latency](<https://devfeed.tech/tags/low-latency.md>), [model](<https://devfeed.tech/tags/model.md>), [open-source](<https://devfeed.tech/tags/open-source.md>), [pipeline](<https://devfeed.tech/tags/pipeline.md>), [ranking](<https://devfeed.tech/tags/ranking.md>), [real-time](<https://devfeed.tech/tags/real-time.md>), [retrieval](<https://devfeed.tech/tags/retrieval.md>), [search](<https://devfeed.tech/tags/search.md>), [sequence](<https://devfeed.tech/tags/sequence.md>), [software-engineering](<https://devfeed.tech/tags/software-engineering.md>)

### AI overview

Airbnb describes extending its sequence recommender with Chronon's Push Mode and Near-real-time Model Transform capabilities. The changes reduce embedding staleness by incorporating recent guest activity instead of waiting for the next batch run.

### Source excerpt

How two new Chronon capabilities, Push Mode and NRT Model Transform, allows us to provide more relevant search results instantly as a guest explores, rather than waiting for the next batch run.

## Beyond the model: Engineering AI infra with scientific judgement

DevFeed: [Beyond the model: Engineering AI infra with scientific judgement](<https://devfeed.tech/articles/beyond-the-model-engineering-ai-infra-with-scientific-judgement-56653.md>)

Original publisher: [Read original article](<https://airbnb.tech/ai-ml/beyond-the-model-engineering-ai-infra-with-scientific-judgement/>)

Author: laurenmackevich

Published: 2026-09-15T22:30:55Z

Content type: article

Language: en

Sources: [Airbnb Engineering & Data Science](<https://devfeed.tech/sources/airbnb-engineering-data-science.md>)

Topics: [Agent Harness](<https://devfeed.tech/topics/agent-harness.md>), [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [Data Science](<https://devfeed.tech/topics/data-science.md>), [data](<https://devfeed.tech/topics/data.md>), [audit](<https://devfeed.tech/topics/audit.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [Large Language Model](<https://devfeed.tech/topics/llm.md>), [trust](<https://devfeed.tech/topics/trust.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agent-harness](<https://devfeed.tech/tags/agent-harness.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-ml](<https://devfeed.tech/tags/ai-ml.md>), [audit](<https://devfeed.tech/tags/audit.md>), [build](<https://devfeed.tech/tags/build.md>), [data](<https://devfeed.tech/tags/data.md>), [data-science](<https://devfeed.tech/tags/data-science.md>), [datasets](<https://devfeed.tech/tags/datasets.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [process](<https://devfeed.tech/tags/process.md>), [products](<https://devfeed.tech/tags/products.md>), [quality](<https://devfeed.tech/tags/quality.md>), [real-world](<https://devfeed.tech/tags/real-world.md>), [trust](<https://devfeed.tech/tags/trust.md>), [unstructured-data](<https://devfeed.tech/tags/unstructured-data.md>)

### AI overview

Airbnb describes Insight Miner, an agent harness for unstructured text exploration. It encodes scientific methodology around an AI model so investigations can be reproducible, auditable, inspectable, and scalable 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.

## When to Retrain, Rebuild, or Leave a Forecasting Model Alone

DevFeed: [When to Retrain, Rebuild, or Leave a Forecasting Model Alone](<https://devfeed.tech/articles/how-we-knew-covid-was-over-and-what-our-models-had-to-unlearn-56656.md>)

Original publisher: [Read original article](<https://airbnb.tech/ai-ml/how-we-knew-covid-was-over-and-what-our-models-had-to-unlearn/>)

Author: laurenmackevich

Published: 2026-08-20T19:33:16Z

Content type: article

Language: en

Sources: [Airbnb Engineering & Data Science](<https://devfeed.tech/sources/airbnb-engineering-data-science.md>)

Topics: [Data Science](<https://devfeed.tech/topics/data-science.md>), [data](<https://devfeed.tech/topics/data.md>), [Risk](<https://devfeed.tech/topics/risk.md>), [structure](<https://devfeed.tech/topics/structure.md>)

Tags: [ai-ml](<https://devfeed.tech/tags/ai-ml.md>), [data](<https://devfeed.tech/tags/data.md>), [data-science](<https://devfeed.tech/tags/data-science.md>), [risk](<https://devfeed.tech/tags/risk.md>)

### AI overview

Airbnb's Forecasting Data Science team explains how it decides whether a struggling forecast needs newer data, a different model, or no change. The article distinguishes refitting, respecifying, and holding a model, while emphasizing validation and production risk.

### Source excerpt

When we retrain, when we rebuild, and when we leave a model alone.

## Flexible Authentication: Reimagining authentication for millions of users at Airbnb

DevFeed: [Flexible Authentication: Reimagining authentication for millions of users at Airbnb](<https://devfeed.tech/articles/flexible-authentication-reimagining-authentication-for-millions-of-users-at-airbnb-56661.md>)

Original publisher: [Read original article](<https://airbnb.tech/uncategorized/flexible-authentication-reimagining-authentication-for-millions-of-users-at-airbnb/>)

Author: laurenmackevich

Published: 2026-08-12T19:03:23Z

Content type: article

Language: en

Sources: [Airbnb Engineering & Data Science](<https://devfeed.tech/sources/airbnb-engineering-data-science.md>)

Topics: [Authentication](<https://devfeed.tech/topics/authentication.md>), [Users](<https://devfeed.tech/topics/users.md>), [account](<https://devfeed.tech/topics/account.md>), [sessions](<https://devfeed.tech/topics/sessions.md>)

Tags: [architecture](<https://devfeed.tech/tags/architecture.md>), [authentication](<https://devfeed.tech/tags/authentication.md>), [login](<https://devfeed.tech/tags/login.md>), [product](<https://devfeed.tech/tags/product.md>), [uncategorized](<https://devfeed.tech/tags/uncategorized.md>), [users](<https://devfeed.tech/tags/users.md>)

### AI overview

Airbnb rebuilt its authentication flows around a two-stage "Identify first then Challenge" model. The system first identifies the account, then uses session context and policy rules to select the most suitable login challenge, with fallback options.

### Source excerpt

Rebuilding login and signup surfaced product insights, not just technical challenges. Here's how we designed Flexible Authentication at the intersection of product intuition and technical architecture.

## Eval-driven development: Lessons from evaluating GenAI at scale

DevFeed: [Eval-driven development: Lessons from evaluating GenAI at scale](<https://devfeed.tech/articles/eval-driven-development-lessons-from-evaluating-genai-at-scale-56654.md>)

Original publisher: [Read original article](<https://airbnb.tech/ai-ml/eval-driven-development-lessons-from-evaluating-genai-at-scale/>)

Author: laurenmackevich

Published: 2026-07-30T17:27:06Z

Content type: article

Language: en

Sources: [Airbnb Engineering & Data Science](<https://devfeed.tech/sources/airbnb-engineering-data-science.md>)

Topics: [Generative AI](<https://devfeed.tech/topics/generative-ai.md>), [Large Language Model](<https://devfeed.tech/topics/llm.md>), [Software Testing](<https://devfeed.tech/topics/software-testing.md>), [Software Engineering](<https://devfeed.tech/topics/software-engineering.md>), [data](<https://devfeed.tech/topics/data.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [ai-ml](<https://devfeed.tech/tags/ai-ml.md>), [development](<https://devfeed.tech/tags/development.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [genai](<https://devfeed.tech/tags/genai.md>)

### AI overview

Airbnb describes evaluation-driven development for trustworthy generative AI products. The article explains why evaluating LLM systems is difficult, including nondeterministic outputs, subjective correctness, chained retrieval and tool calls, and the risk of evaluator failure. It recommends planning evaluation early, reviewing data manually, and measuring dimensions that reflect real user outcomes.

### Source excerpt

How Airbnb teams build trustworthy Generative AI products by treating evaluation as a first-class engineering discipline; not an afterthought.

## Personalizing Airbnb search by learning from the guest journey

DevFeed: [Personalizing Airbnb search by learning from the guest journey](<https://devfeed.tech/articles/personalizing-airbnb-search-by-learning-from-the-guest-journey-56657.md>)

Original publisher: [Read original article](<https://airbnb.tech/ai-ml/personalizing-airbnb-search-by-learning-from-the-guest-journey/>)

Author: laurenmackevich

Published: 2026-07-27T19:00:17Z

Content type: article

Language: en

Sources: [Airbnb Engineering & Data Science](<https://devfeed.tech/sources/airbnb-engineering-data-science.md>)

Topics: [recommendation systems](<https://devfeed.tech/topics/recommendation-systems.md>), [Sequences](<https://devfeed.tech/topics/sequences.md>), [Transformer](<https://devfeed.tech/topics/transformer.md>), [Model Development](<https://devfeed.tech/topics/model-development.md>)

Tags: [ai-ml](<https://devfeed.tech/tags/ai-ml.md>), [optimizations](<https://devfeed.tech/tags/optimizations.md>), [ranking](<https://devfeed.tech/tags/ranking.md>), [sequence](<https://devfeed.tech/tags/sequence.md>), [transformer](<https://devfeed.tech/tags/transformer.md>)

### AI overview

Airbnb describes a Transformer-based sequence modeling system that learns from years of guest behavior, including views, bookings, reviews, and cancellations, to improve personalized search ranking. The article explains challenges involving long, noisy event sequences, sparse booking signals, and the cost of training at scale.

### Source excerpt

How we built a Transformer-based sequence model that encodes years of guest behavior to surface the right listings at the right time.

## From weeks to a day: how we made LLM evaluation fast enough to iterate on

DevFeed: [From weeks to a day: how we made LLM evaluation fast enough to iterate on](<https://devfeed.tech/articles/from-weeks-to-a-day-how-we-made-llm-evaluation-fast-enough-to-iterate-on-56655.md>)

Original publisher: [Read original article](<https://airbnb.tech/ai-ml/from-weeks-to-a-day-how-we-made-llm-evaluation-fast-enough-to-iterate-on/>)

Author: laurenmackevich

Published: 2026-07-27T18:57:46Z

Content type: article

Language: en

Sources: [Airbnb Engineering & Data Science](<https://devfeed.tech/sources/airbnb-engineering-data-science.md>)

Topics: [LLM evaluation / benchmarking](<https://devfeed.tech/topics/llm-evaluation-benchmarking.md>), [Large Language Model](<https://devfeed.tech/topics/llm.md>), [Software Engineering](<https://devfeed.tech/topics/software-engineering.md>), [systems](<https://devfeed.tech/topics/systems.md>), [Machine learning](<https://devfeed.tech/topics/machine-learning.md>)

Tags: [ai-ml](<https://devfeed.tech/tags/ai-ml.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [llm](<https://devfeed.tech/tags/llm.md>), [llm-evaluation](<https://devfeed.tech/tags/llm-evaluation.md>), [quality](<https://devfeed.tech/tags/quality.md>), [validation](<https://devfeed.tech/tags/validation.md>)

### AI overview

Airbnb describes the infrastructure and evaluation techniques it used to reduce production LLM evaluation time from weeks to a day. The article focuses on trustworthy measurement, evaluation noise, end-to-end validation, and integrating multiple layers of the production LLM stack.

### Source excerpt

Training an LLM is the easy part. The hard part is designing experiments and evaluations that you can trust enough to know whether the new model is actually an improvement.

## Scaling beyond one: How Airbnb evolved its data architecture for a multi-product world

DevFeed: [Scaling beyond one: How Airbnb evolved its data architecture for a multi-product world](<https://devfeed.tech/articles/scaling-beyond-one-how-airbnb-evolved-its-data-architecture-for-a-multi-product-world-56662.md>)

Original publisher: [Read original article](<https://airbnb.tech/uncategorized/scaling-beyond-one-how-airbnb-evolved-its-data-architecture-for-a-multi-product-world/>)

Author: laurenmackevich

Published: 2026-06-09T19:34:04Z

Content type: article

Language: en

Sources: [Airbnb Engineering & Data Science](<https://devfeed.tech/sources/airbnb-engineering-data-science.md>)

Topics: [data-architecture](<https://devfeed.tech/topics/data-architecture.md>), [data](<https://devfeed.tech/topics/data.md>), [data-modeling](<https://devfeed.tech/topics/data-modeling.md>), [Data Infrastructure](<https://devfeed.tech/topics/data-infrastructure.md>), [systems](<https://devfeed.tech/topics/systems.md>)

Tags: [2025](<https://devfeed.tech/tags/2025.md>), [airbnb](<https://devfeed.tech/tags/airbnb.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [data](<https://devfeed.tech/tags/data.md>), [data-architecture](<https://devfeed.tech/tags/data-architecture.md>), [data-engineers](<https://devfeed.tech/tags/data-engineers.md>), [data-infrastructure](<https://devfeed.tech/tags/data-infrastructure.md>), [data-modeling](<https://devfeed.tech/tags/data-modeling.md>), [data-systems](<https://devfeed.tech/tags/data-systems.md>), [flexible](<https://devfeed.tech/tags/flexible.md>), [new-product](<https://devfeed.tech/tags/new-product.md>), [scaling](<https://devfeed.tech/tags/scaling.md>), [services](<https://devfeed.tech/tags/services.md>), [technical](<https://devfeed.tech/tags/technical.md>), [technical-debt](<https://devfeed.tech/tags/technical-debt.md>), [uncategorized](<https://devfeed.tech/tags/uncategorized.md>)

### AI overview

Airbnb's data and analytics engineers evolved its offline data warehouse and data-modeling framework to support Homes, Experiences, and Services. The article examines the trade-offs between separate product-specific models and a unified monolithic model.

### Source excerpt

How Airbnb's data engineers and analytics engineers built a consistent and flexible data modeling framework to support the expansion into Homes, Experiences, and Services.

## Sitar-agent: Building a reliable dynamic configuration sidecar at scale

DevFeed: [Sitar-agent: Building a reliable dynamic configuration sidecar at scale](<https://devfeed.tech/articles/sitar-agent-building-a-reliable-dynamic-configuration-sidecar-at-scale-56660.md>)

Original publisher: [Read original article](<https://airbnb.tech/infrastructure/sitar-agent-building-a-reliable-dynamic-configuration-sidecar-at-scale/>)

Author: laurenmackevich

Published: 2026-06-04T21:06:52Z

Content type: article

Language: en

Sources: [Airbnb Engineering & Data Science](<https://devfeed.tech/sources/airbnb-engineering-data-science.md>)

Topics: [configuration](<https://devfeed.tech/topics/configuration.md>), [Kubernetes](<https://devfeed.tech/topics/kubernetes.md>), [Amazon S3](<https://devfeed.tech/topics/amazon-s3.md>), [container](<https://devfeed.tech/topics/container.md>), [Amazon Web Services](<https://devfeed.tech/topics/aws.md>), [Git](<https://devfeed.tech/topics/git.md>), [versioning](<https://devfeed.tech/topics/versioning.md>), [service](<https://devfeed.tech/topics/service.md>), [Back end](<https://devfeed.tech/topics/backend.md>)

Tags: [aws](<https://devfeed.tech/tags/aws.md>), [aws-s3](<https://devfeed.tech/tags/aws-s3.md>), [backend](<https://devfeed.tech/tags/backend.md>), [configuration](<https://devfeed.tech/tags/configuration.md>), [container](<https://devfeed.tech/tags/container.md>), [deployments](<https://devfeed.tech/tags/deployments.md>), [developer](<https://devfeed.tech/tags/developer.md>), [git](<https://devfeed.tech/tags/git.md>), [infrastructure](<https://devfeed.tech/tags/infrastructure.md>), [kubernetes](<https://devfeed.tech/tags/kubernetes.md>), [scale](<https://devfeed.tech/tags/scale.md>), [sidecar](<https://devfeed.tech/tags/sidecar.md>)

### AI overview

Airbnb describes how Sitar-agent, a lightweight Kubernetes sidecar, synchronizes dynamic configuration to subscribed service pods. It uses versioned configuration changes, AWS S3 snapshots, startup synchronization, and periodic polling to provide reliable updates without redeploying services.

### Source excerpt

How Airbnb built a Kubernetes sidecar to deliver dynamic configuration reliably at scale.

## Airbnb Uses Geographic Recovery Signals to Forecast Demand When Local Data Is Scarce

DevFeed: [Airbnb Uses Geographic Recovery Signals to Forecast Demand When Local Data Is Scarce](<https://devfeed.tech/articles/when-history-fails-you-borrow-from-geography-56658.md>)

Original publisher: [Read original article](<https://airbnb.tech/data/when-history-fails-you-borrow-from-geography/>)

Author: laurenmackevich

Published: 2026-06-03T22:19:25Z

Content type: article

Language: en

Sources: [Airbnb Engineering & Data Science](<https://devfeed.tech/sources/airbnb-engineering-data-science.md>)

Topics: [data](<https://devfeed.tech/topics/data.md>), [standard](<https://devfeed.tech/topics/standard.md>)

Tags: [airbnb](<https://devfeed.tech/tags/airbnb.md>), [covid](<https://devfeed.tech/tags/covid.md>), [data](<https://devfeed.tech/tags/data.md>), [forecasting](<https://devfeed.tech/tags/forecasting.md>), [geography](<https://devfeed.tech/tags/geography.md>), [model](<https://devfeed.tech/tags/model.md>), [patterns](<https://devfeed.tech/tags/patterns.md>)

### AI overview

Airbnb describes a forecasting approach for the COVID recovery period that used sequential signals from other geographies and prior propagation when individual markets lacked sufficient post-shock data. The method produced corridor-level forecasts across uneven reopening and reclosure patterns.

### Source excerpt

How Airbnb used sequential geographic recovery signals and prior propagation to generate reliable corridor-level forecasts when local data was scarce.

## Scaling Airbnb's identity graph with a unified knowledge graph infrastructure

DevFeed: [Scaling Airbnb's identity graph with a unified knowledge graph infrastructure](<https://devfeed.tech/articles/scaling-airbnb-s-identity-graph-with-a-unified-knowledge-graph-infrastructure-56659.md>)

Original publisher: [Read original article](<https://airbnb.tech/infrastructure/scaling-airbnbs-identity-graph-with-a-unified-knowledge-graph-infrastructure/>)

Author: laurenmackevich

Published: 2026-05-19T17:51:47Z

Content type: article

Language: en

Sources: [Airbnb Engineering & Data Science](<https://devfeed.tech/sources/airbnb-engineering-data-science.md>)

Topics: [graph-database](<https://devfeed.tech/topics/graph-database.md>), [Graphs](<https://devfeed.tech/topics/graphs.md>), [Scalability](<https://devfeed.tech/topics/scalability.md>), [Database](<https://devfeed.tech/topics/database.md>), [trust and safety](<https://devfeed.tech/topics/trust-and-safety.md>), [systems](<https://devfeed.tech/topics/systems.md>)

Tags: [architecture](<https://devfeed.tech/tags/architecture.md>), [caching](<https://devfeed.tech/tags/caching.md>), [database](<https://devfeed.tech/tags/database.md>), [graph](<https://devfeed.tech/tags/graph.md>), [infrastructure](<https://devfeed.tech/tags/infrastructure.md>), [json](<https://devfeed.tech/tags/json.md>), [knowledge-graph](<https://devfeed.tech/tags/knowledge-graph.md>), [latency](<https://devfeed.tech/tags/latency.md>), [performance](<https://devfeed.tech/tags/performance.md>), [real-time](<https://devfeed.tech/tags/real-time.md>), [saas](<https://devfeed.tech/tags/saas.md>), [scaling](<https://devfeed.tech/tags/scaling.md>)

### AI overview

Airbnb describes the evolution of its identity graph and its move to an internally managed unified knowledge graph infrastructure. The article covers the graph's architecture, storage and service layers, scaling challenges, performance issues, and optimizations.

### Source excerpt

How Airbnb shifts from PaaS to an internal knowledge graph infrastructure at scale.