I'm Muhammad Usman Khalil, an AI Engineer and Full-Stack Developer with 4 years of experience building production AI systems, agents, and automation. I help companies turn AI capabilities into real business outcomes.

I'm an AI engineer with 4 years of experience building production-grade ML and agentic AI systems. My focus is the applied LLM stack: RAG pipelines, LLM fine-tuning for domain-specific tasks, and multi-agent workflows using LangGraph, CrewAI, and the Model Context Protocol (MCP). I've also shipped voice agents and intelligent chatbots that handle real customer conversations end to end.

Day to day, I work with Python, PyTorch, Hugging Face, and LangChain, paired with vector databases like Qdrant and Pinecone for retrieval. I take systems all the way to production: containerizing services with Docker, building CI/CD pipelines, and deploying to AWS and Azure with an eye on cost and latency.

I'm also a full-stack developer (React, Next.js, Node.js), which means I can wrap the models I build into actual products people use instead of handing off a notebook and walking away.

The way I work is consistent and disciplined. I show up every day, ship in small steady increments rather than big unreliable bursts, and I'm always trying to learn the next thing, whether that's a new framework, a new architecture pattern, or a sharper way to think about a problem. Most of what I enjoy is the figuring-out part: taking a vague problem, breaking it down, and getting to a clean solution.

What sets me apart isn't any one tool in the stack. It's that I treat AI work as engineering, not magic. I care about whether the system actually runs in production, whether it stays cheap to operate, and whether the people using it get real value back. Plenty of people can build a demo. Fewer can ship something that holds up.

When I'm not coding, I enjoy playing video games and gardening. I'm also learning history and finance, and picking up how to build businesses on the side.