Tool Launches

Meta Deploys JiTTesting to Catch Code Bugs at Scale

Source: Meta AI (Engineering) · Published Oct 29, 2026 · Summarised Nov 1, 2026

Read the original on Meta AI (Engineering)
AI-generated summary. This summary was written by AI (Google Gemini) using only the text Meta AI (Engineering) published in its feed. It may omit details; read the original article for the full story.

Meta has successfully deployed its Just-in-Time Tests approach into production to catch software bugs before code lands in production. The system utilizes large language models to generate executable tests that demonstrate concrete behavioral differences introduced by code changes, addressing the limitations of traditional testing methods on massive codebases.

The deployment relies on two primary workflows: an intent-aware workflow that infers code risks and a dodgy diff workflow that treats code changes as potential mutants. To minimize false positives, the infrastructure combines an LLM-as-Judge ensemble with a traditional rule-based assessor called RubFake. This automated filtering is paired with a human-on-the-loop approach, reducing the review load on engineers while maintaining high true-positive rates for bug detection.

Key takeaways

  • Meta deployed Just-in-Time Tests into production to catch code bugs using large language models.
  • The system uses intent-aware and dodgy diff workflows alongside an LLM-as-Judge ensemble to filter false positives.
  • A human-on-the-loop framework helps reduce review loads while ensuring reliable bug detection at scale.

Original source: Meta AI (Engineering). TryAITools is not affiliated with Meta AI (Engineering); all trademarks belong to their owners.

← All AI News