Curriculum

Our field is moving fast, and this curriculum is designed to move with it. We will pivot modules, swap technologies, and adjust timelines based on what we believe is most useful learning.

What this curriculum provides is a blueprint: the overarching themes, the progression of skills, and a sense of where we are heading. We use it as a steer and a yardstick, but we will not let it get in the way of improvements that can be made along the way.

Programme overview

The programme is structured around ten modules. Most modules centre on a project: you build something real and ship it. Others focus on foundational skills or operational practices that support the rest of the programme. Modules are sequenced so that skills and infrastructure carry forward.

Working with coding agents

You will use agentic coding tools from the first week of this programme. This is a craft skill that develops over time. In practice it means learning to structure tasks for delegation, writing project context files that give agents what they need, knowing when to intervene, and reviewing AI-generated code with the same rigour you would apply to a pull request from a colleague. As a cohort, you will share what works and what doesnt, and build a collective understanding of how to work effectively with agents. Things are changing from week to week, and together we will be exploring new ways of working.

Skills you will gain

TransformersAutonomous agentsProduction deploymentMLOpsMulti-agent systemsModel servingModel routingDockerKubernetesPyTorchReal-time voice AIBackpropagationAttention mechanismsFine-tuning & LoRAQuantizationMCP serversRed teamingPrompt injection defenseRAG systemsVector searchContext engineeringDevOpsSynthetic dataEvaluation & testingClassical MLscikit-learn & XGBoostLinear algebraCalculus & probability

Modules