Towards an open operations data community
summary
AI4CI collects open CI/CD data from Prow, TestGrid and GitHub, stores it on Ceph S3, and runs Jupyter notebooks automated with Kubeflow Pipelines to compute KPIs and populate Superset dashboards. It deploys trained models with the Seldon operator on OpenShift — example services include a time-to-merge model that classifies PRs into 10 buckets (roughly 20–25% accuracy per bucket with vanilla classifiers) and an optimal-test-stop model that returns points like 104 seconds to abort long runs. All code, data and pipelines are public.
AI-written summary · about
chapters (9)
- 0:00Introduction to open ops data
- 4:39Why open operations data matters
- 7:28AI for CI project overview
- 11:18Demo time to merge model
- 16:13Demo optimal test stopping
- 18:01Automating ML workflows
- 21:11Superset dashboard walkthrough
- 23:18How to engage with the project
- 24:40Q&A session
related talks
24mUncovering Community and Project Insights Through Data Driven Methods
44mWhat’s Breaking My Build? AIOps Tooling to Aid Your CI/CD Workflows
38mBI DataOps Development Process
44mAgentic Delivery, Guardrailed: AI in CI/CD & Platform Engineering (Open Source Patterns)
21mAI-driven optimization of CI/CD pipelines
05mData Science in DevOps/SysOps
34mData pipelines powered by Open source and fun
20mBuild an e2e analytics application using DataHub