I build self-enhancing AI applications and autonomous workflows that empower people to do more with less friction.
- 🧠 Intent Engineering — I design AI systems that understand what you mean, not just what you say. My work focuses on building agents that learn, adapt, and act with increasing autonomy.
- 🍄 Myco — Persistent cognitive layer for AI agents. Gives Claude Code sessions long-term episodic memory, semantic search via local Ollama embeddings, and a knowledge graph. Includes a React PWA dashboard.
- 🔍 Fathom — Intelligence layer that makes AI coding costs intentional and trackable. Estimates token usage per feature, tracks actual spend across providers, and self-calibrates over time. CLI + MCP server + SDK.
- 📈 TradeBot — Autonomous swing trading system that scans markets, paper-trades with disciplined risk management, and learns from its own history via local embeddings. Graduating toward live trading.
- 🗺️ Voya — Collaborative trip planning with AI-powered suggestions, real-time collaboration, and drag-and-drop itinerary building.
- Building autonomous AI agents that self-learn and self-improve
- Local-first AI — Ollama embeddings, vector stores, keeping data on-device
- MCP servers and composable AI tooling
- Intent engineering — designing prompts and workflows that translate human goals into reliable agent behavior
- Closing the loop: scan → decide → act → learn → improve
Languages & Frameworks
AI & ML
Data & Infrastructure
