Transforming Software Development and Engineering Velocity at EBay With AI/ML in C...
summary
eBay describes an internal AI/ML CI/CD platform that measurably speeds delivery: 1,600+ daily commits across 4,000 apps, commit‑to‑prod down to ~45 minutes, lead time improving by 11 hours year‑over‑year and developer friction falling one‑third in six months. Built components include a configurable code review agent with ~40% acceptance and contributing to a 10% drop in production bugs, an ECDX agent that runs RCA and chat flows against logs and CI events pulling from a vector database, and automated pipeline optimizations saving ~20 minutes across 500 apps.
AI-written summary · about
chapters (13)
- 0:00AI-driven engineering velocity
- 1:26Developer friction and DORA metrics
- 3:13AI in inner development loop
- 4:15PR summarization and code review
- 5:21RCA for test failures and pipelines
- 6:43Developer friction dashboard and AI
- 8:27AI insights and pipeline optimization
- 9:53AI support widget and troubleshooting
- 12:00RAG responses and Slack AI agents
- 13:46ECDX agent architecture
- 16:39AI adoption results and impact
- 17:41Learnings and best practices
- 19:28Future roadmap and pipeline factory
related talks
45mImproving eBay's Development Velocity
36mDoubling Engineering Productivity at eBay Through DevOps
26mAccelerating DevOps with AI
29mWe doubled engineering productivity at eBay, but couldn't change culture
47mAI, Automation & DevEx: Fueling High-Velocity Engineering
44mSimplifying and Scaling Software Delivery
27mFrom AI Assistants to AI Coworkers: How Engineering Teams Ship Faster with Enterprise Context
55mAI Assisted Engineering