Reproduce “Automating Feature Flag Cleanup at Scale with Multi-Agent LLM Systems: An Industrial Experience at DoorDash” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Atharv Chandratre, Jai Datar
The system removes stale DoorDash Dynamic Values by checking each flag’s live rollout state, having an engineer confirm its target value, then using isolated LLM agents to edit references and tests and run validation gates. Across 50 Kotlin-repository cases, it produced usable pull requests for 45 (90%), including revisions; five required engineer intervention, especially in observed deep call chains.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.