Building secure AI/ML systems that make multi-agent software cool and reliable in production. Agent-native architecture - data models, execution pipelines, comm protocols, safety boundaries. Deep on alignment and security design for AI that don't go rogue.
Main stack is Python, Go, PyTorch, FastAPI, PostgreSQL, Bubble Tea and the Charm ecosystem for TUIs, and LLM application architecture. Currently pushing hard on Go (TUI + backend).
Beyond the Code
Exercise science and gym outside the terminal. Music has been a main hobby and these are the records on repeat:
