AI Engineer. I build production GenAI systems: agents, RAG pipelines, and the infrastructure around them.
I build GenAI products that run in production, not in notebooks. Two systems I have shipped:
- Auxee β an enterprise GenAI chatbot on Azure (Azure OpenAI, Azure AI Search, RAG over private company data)
- Falconic AI β an autonomous software engineering platform built on LangGraph agents with sandboxed code execution (E2B)
Most of my public repos here are focused explorations of the same ideas: multi-agent supervision, agentic RAG, MCP clients and servers, and production-grade LangGraph patterns.
Currently going deeper on evals (DeepEval, PromptFoo) and observability (LangSmith), because agents that can't be measured can't be trusted.
| Repo | What it shows |
|---|---|
| Production-Grade-LangGraph-Chatbot-Agent | Structuring a LangGraph agent the way you would for a real deployment |
| Supervisor-Multi-Agent | Supervisor pattern for coordinating multiple agents |
| Agentic-Rag | RAG where the retrieval strategy is decided by the agent, not hardcoded |
| Custome-MCP-Client-Server | MCP client and server built from scratch to understand the protocol |
I write about agent architectures and RAG patterns on LinkedIn. If you are working on agentic systems and want to compare notes, reach out.