Projects

• guard
  Security scanner for AI applications — catches leaked API keys and prompt injection before they reach production.
  GitHub
• measure
  Production LLM evaluation dashboard that turns "my LLM works great" into real, trackable metrics.
  GitHub
• router
  Intelligent LLM routing that classifies query complexity and routes accordingly, cutting cost by up to 60%.
  GitHub
• retrieval
  Local-first RAG pipeline where every answer cites its source page — grounded responses instead of hallucinations.
  GitHub