Oyediji Yinka

I build and ship AI systems that solve real problems and show real results. Most of my work is agentic—RAG pipelines, LangGraph agents, and FastAPI backends—deployed with Docker on AWS, with results like workflows around 90% more efficient and customer churn cut by up to 80%.

I’m comfortable across the stack: Python and PyTorch, plus PostgreSQL, Neo4j, and Redis. I came into AI from mining engineering, so I’m used to turning messy real-world data into something people can actually use, and I work well in remote, distributed teams.

In my current AI / Backend role, I build production AI agents end to end—from the agent loop and RAG retrieval to evaluation and deployment. I’ve shipped a RAG email agent with human-in-the-loop approval (around 90% more efficient email handling), an AI customer support agent that cuts churn by up to 80%, and a voice assistant that pairs RAG with speech-to-text and text-to-speech.