I build backend systems and full-stack apps, mostly in TypeScript and Python.
- PR Review & Regression Testing Platform: AI code review that cites exact changed lines and proves suspected bugs with failing tests before proposing a fix. Benchmarked on 50 real PRs against 17 commercial reviewers. Results
- Agent Reliability & Evaluation Platform: Gates agent releases on repeated trials, verified with hidden tests and independent state checks. Evidence explorer
- Global Macro Research Workbench: Live research site that refreshes FRED, BIS, World Bank and market data every 12 hours. Live site