Building governed, reliable, and measurable AI agents — not demos.
Teaching to ship AI-native and AI-first products through specification, not syntax. Faculty at Panaversity and PIAIC. Taught Python at Stanford CIP in 2024.
Methodology: Spec-Driven Development with Reusable Intelligence (SDD-RI)
Open source:
- SpecKit Plus — SDD-RI framework
- AI-Native Book — co-authored by agents and humans
- RoboLearn — AI-native textbook platform (built in 48 hours using SDD-RI)
- TaskFlow — AI-first project management for Agents (Claude Code) and Humans.
- Architecture: Azure AKS, Dapr Sidecars, 5 Micro-services, and Neon PostgreSQL.
- Agentic: OAuth 2.0 Compliant MCP Server with ChatKit Gen-UI.
- Methodology: Fully Spec-Driven (19 specs) with 250+ CI tests and Zero-touch CD.
Languages: Python · TypeScript
Stack: Claude Code · Gemini CLI · MCP · Docker · Kubernetes · Dapr · A2A
Agent Frameworks: OpenAI Agents SDK · Google ADK · LangGraph
Frontend: Next.js · ChatKit




