01
Large language models
What a model can do, what it cannot, and how to place one inside a process that already works.
- 01 Prompting as an interface contract
- 02 Retrieval and grounding
- 03 Fine-tuning: when it pays
- 04 Cost and latency budgets
The catalogue
Four tracks, one standard.
Every programme is written by someone who does the job they teach, and rewritten after each cohort.
02
Applied data analysis
From a question someone actually asked to a number a decision can rest on.
- 01 SQL beyond SELECT
- 02 Modelling for decisions
- 03 Applied statistics
- 04 Reporting that gets read
03
Systems and deployment
The unglamorous half: serving, queues, timeouts, cost and the pager at 3am.
- 01 Serving and latency
- 02 Evaluation harnesses
- 03 Observability in production
- 04 Incident response
04
AI product design
Research, prototyping, testing and handover to the engineers who build it.
- 01 Problem framing
- 02 Prototyping with models
- 03 User testing under uncertainty
- 04 Handover artefacts
January 2027 cohort
Not sure which track?
The scoping call exists to answer exactly that, and it costs nothing.
Book a scoping call