/developer-experience-ai
Faster delivery without lower quality
I build the tools and standards that make good engineering the default.
Architectural invariants
- Conventions live in code and machine-checkable rules, not wikis.
- Design happens before code, and a template keeps it under two hours.
- Pipelines are composed from shared building blocks, never copy-pasted.
- AI reviewers check the rules so humans can review the design.
Proof in production
- Created a “deliver faster without compromising quality” workflow that cut technical design doc time from 2 days to 2 hours.
- Built a library of reusable GitHub Actions adopted across engineering, data science and data engineering.
- Wrote
AGENTS.mdcoding rules that AI review agents cite in pull request comments. - Raised test coverage 25% and halved failed weekend health-check alerts.
- Built new-hire laptop setup automation and a reference service that shows the team's patterns.