Software Engineering for AI Systems
Four weeks on the engineering practice AI work assumes but rarely teaches: version control that survives collaboration, APIs that fail well, containers, CI, and the observability you will need the first time something breaks in production.
Join the waitlistWho should take this program?
You'll get the most from this if
- Able to write and run Python
- Basic command line familiarity
This is not for
- Senior engineers who already ship containerised services with CI
What you will build
- A containerised service with CI running tests on every push
- An API with proper error contracts and structured logs
- A runbook someone else could follow at 2am
What you learn, week by week
Module 1 — Git as a team tool
- Branch, review and merge without fear
- Write commits someone can read later
- Recover from the usual disasters
Module 2 — APIs that fail well
- Design clear request and response contracts
- Return errors a client can act on
- Version an API without breaking callers
Module 3 — Containers and environments
- Containerise a Python service
- Keep images small
- Manage secrets without committing them
Module 4 — CI and observability
- Run tests automatically on every push
- Emit structured logs and traces
- Write a runbook for the failure you expect
How this program handles evaluation and governance
Auditability is an engineering property before it is a policy one. If your service has no structured logs, no CI and no runbook, no governance document will make its behaviour reconstructable. This course builds that substrate.
Tools and stack you will use
Fees
Request pricing
Fees depend on cohort, format, and whether this runs for an individual or a team. We'll send the full breakdown.
Get fee detailsQuestions people ask
Is this DevOps?
A slice of it — the slice AI engineers actually need. Not Kubernetes administration.
Can I skip this?
If you already ship production services, yes. If your experience is notebooks, do not.