Kipo AI
| San FranciscoJun 2025 - Present
Software Engineering Intern
Built and hardened core platform systems as the product scaled from stealth to ~1,000 active users with 2M+ indexed components.
- Rewrote the authentication and session system end-to-end to unblock OAuth and enterprise rollout, with cross-browser end-to-end tests guarding the flow.
- Made the AI agent reliably usable for ~1,000 users by introducing a durable workflow engine and moving chat state onto Redis, eliminating the race conditions that were dropping live responses.
- Led an architectural split that decoupled the chat and search subsystems, and hardened the deploy pipeline with reproducible builds, faster caching, and end-to-end test coverage.
Languages:PythonTypescript
Frontend:Next.js
Backend:NestJSFastAPIOpenFGA
Databases:PostgreSQLMongoDBRedis
Services:AWS