Curriculum

Our programme is built around weekly practical workshops where you'll work on real engineering challenges. Rather than following a fixed curriculum, sessions respond to the rapidly evolving AI landscape and what participants are actually building. You'll tackle problems like integrating semantic search into existing applications, building multi-agent workflows for production systems, and implementing RAG architectures that handle real-world constraints around cost and latency.

Each week combines hands-on implementation with collaborative problem-solving. You might spend one session building an MCP server to extend Claude's capabilities, another optimising vector search performance, and another fine-tuning open-source models for specific use cases. The focus is on shipping working features, not academic exercises.

The technical content spans the full stack of modern AI engineering: from using frontier models effectively through APIs, to understanding transformers well enough to customise them for production. You'll work with PyTorch to build neural networks from first principles, implement attention mechanisms, and gain the foundation needed to make informed decisions about model selection, fine-tuning, and deployment. When you need to containerise models, manage data pipelines, or handle GPU orchestration, you'll learn the MLOps practices that make systems reliable at scale.