You have a product. I add the intelligence layer that makes it autonomous — AI agents, chatbots, document Q&A, semantic search, or automated analysis — built properly, not as a quick hack.

I work with OpenAI (GPT-4o, o3), Anthropic Claude, and open-source models. For knowledge-grounded features, I build RAG pipelines with vector search (pgvector, Pinecone, Weaviate). For autonomous workflows, I build multi-step AI agents using function calling, tool use, and orchestration frameworks (LangChain, LlamaIndex, or custom). I handle streaming responses, prompt engineering, and cost optimisation — all exposed through a clean API your team can maintain.

AI logic stays cleanly separated from your business logic — testable, replaceable, and ready to evolve as models improve.