Build agents that execute, fail loudly, and stay under human control.
Founder building AI B2B SaaS.
I design multi-agent systems that research, reason, collaborate, and execute real business workflows
- Agent orchestration & execution graphs
- Long-running, stateful, failure-aware agents
- Human-in-the-loop systems (approve, override, audit)
- Tool reliability, evals, and infra for LLM systems
- Active contributor to an open-source agent orchestration framework
- Shipped production MCP integrations: Google Maps · Apollo · Cal.com · X
- Built multi-agent B2B workflows (research → approval → execution)
- Worked on agent state handling, retries, evals, and failure modes
(Details live in commits, PRs, and issues.)
AI — Claude · Gemini · OpenAI · LangChain · RAG
Backend — Python · FastAPI · TypeScript
Data — PostgreSQL · MongoDB · Redis
Infra — AWS · GCP · Docker · Kubernetes · Terraform

