I build production AI systems that companies actually depend on — from enterprise LLM pipelines and RAG architectures to intelligent trading advisors. I believe the best AI isn't the flashiest; it's the one that works at 3 AM when no one's watching.
deja — Memory for Claude CodeClaude Code forgets everything between sessions. deja fixes that. A plugin that gives Claude persistent, structured memory using local SQLite — no cloud, no latency, no data leaving your machine.
|
AIUPTO — AI Stock Decision AdvisorAn intelligent trading advisor that combines real-time market data with AI analysis to surface actionable stock decisions. Not another chatbot — a decision engine.
|
IBKR MCP Server — Interactive Brokers × AIAn MCP server that bridges Interactive Brokers with AI agents — enabling autonomous portfolio analysis, trade execution monitoring, and risk assessment through natural language.
|
TrailingMaster — Intelligent Trailing StopsA trailing stop advisor that dynamically adjusts stop-loss levels using volatility analysis and market regime detection. Protects profits without giving back gains.
|
AI & LLM
Languages & Frameworks
Cloud & Infrastructure
Data
"The best AI systems don't replace human judgment — they augment it."
I spend my days at the intersection of AI engineering and real-world systems. Whether it's building RAG pipelines that handle millions of documents, creating AI agents that trade financial markets, or making developer tools smarter — I care about AI that ships, scales, and stays reliable.
Currently exploring: AI agents for quantitative finance, MCP protocol integrations, and developer tooling for the AI era.
If you're building at the intersection of AI and finance, let's talk.
262 repos and counting — because shipping beats planning.