Linear Algebra with Python
Linear systems and matrices, and the vector space structure underlying them, implemented in NumPy.
I build LLM systems that hold up in production: agents, RAG and evals. Open to remote software and AI engineering roles and contract work.

Case studies of systems I designed and built: the architecture, the decisions and what came out of them.

The third engineering case study: how Aluna screens candidates with six specialised agents instead of one chatbot, why e…

The second engineering case study: how CREARIA Agent answers customer questions on WhatsApp using RAG over pgvector and …

A behind-the-scenes case study of Plixiq — why we picked FastAPI, Next.js, LiteLLM and a modular monolith to run AI agen…
Free courses where I teach the math and engineering behind modern AI, from the original formulations.
Linear systems and matrices, and the vector space structure underlying them, implemented in NumPy.
What exactly is optimised when a language model is trained: self-supervised pre-training, the scaling laws and their correction, and the objectives that separate one architecture family from another.
Python and TypeScript on the backend and frontend, PostgreSQL and Redis for data, and LLM providers behind one interface.
Join a weekly study group to explore mathematics, software engineering, and artificial intelligence together.