Building Computer Vision infrastructure & AI-native Analytical tools at Paloa Labs in Palo Alto, CA.
I'm obsessed with making AI agents that actually work — event-driven systems, reinforcement learning from human feedback, and tools that turn messy real-world interfaces into clean APIs.
| Project | What it does |
|---|---|
| 🌐 site2cli | Turn any website into a CLI/API automatically — on PyPI, 156 tests passing |
| 🏎️ aero-agent | CFD-aware research assistant for F1 aerodynamicists — synthesize literature, correlate CFD, optimize every simulation hour under FIA ATR constraints |
| 🧬 FounderAgent | Autonomous health-aware AI agent — reads your biometrics, blocks calendar, sets Slack DND, sends recovery protocols |
| 🏆 FounderLeague | Gamified league platform where founders compete on readiness, sleep quality & recovery — anti-hustle-culture leaderboards |
| 🧠 GameDay-AMS | Cognitive Performance System for Founders — sports science meets knowledge work |
| 🤖 agentic-employees | Event-driven AI agents with GRPO training, NATS messaging, and appraisal-style feedback loops |
| 🔬 MCPstudio | The Postman for MCP — create, test, manage & discover MCP servers with a visual UI. FastAPI + React, DDD architecture, Google Drive integration, JWT auth |
| 🏀 paloa-claw | Personal basketball AI training assistant |
| 🎓 university-sim | Multi-agent university simulation benchmark with Concordia GM + OpenClaw |
| ⚡ RLYX-enhancer | Hackable GRPO training framework with high-speed weight sync for multi-node RL |
| 🩺 AgentKit-Health | Marketplace of pre-built AI agent templates for health workflows — Sleep Coach, Burnout Detector, Recovery Agent & more |
| 🏠 HandyHommieAI | AI-powered home appliance command center — upload manuals, ask questions, auto-call manufacturer support |
| 🃏 blackjack-for-toddler | Toddler-friendly Blackjack that teaches numbers & addition — React PWA with Capacitor for iOS |
| 🃏 poker-for-toddler | Texas Hold'em for kids (ages 4–9) with toy cars, gems & crowns — React PWA on Vercel |
- AI Agent Systems — event-driven architectures, multi-agent coordination, tool use
- Reinforcement Learning — GRPO, DAPO, process reward models, RLHF
- Developer Tools — MCP servers, CLI generators, API discovery
- Applied ML — federated RL (flower.ai contributor), deep finance RL
- Founder Performance — biometric-driven agents, cognitive performance systems, gamified health optimization
- Health & Wellness AI — agent templates for wearable data, sleep coaching, burnout detection
- F1 & Motorsport AI — CFD-aware research assistants, aerodynamic optimization under FIA constraints
- EdTech for Kids — gamified learning apps for toddlers, math through card games
"Ship beats perfect" — I build things that solve real problems, then open-source them.