Season 8: You Own the Release · Episode 1 of 8
Stop Deploying by Memory
You play: AI Engineering Intern
Included with Core programme and Certification prep
The situation
assist-service has outgrown the days when Kabelo and Maya were the only two people who ever touched main. There's a small team on it now, and the habits that were fine solo — push straight to main, merge your own change, run the tests if you remember to — are starting to leave marks nobody agreed to.
What you'll learn
- Branch protection with a required review and a required status check that actually blocks an unready merge; reading the commit/PR audit trail to find who changed what and when; using it to diagnose and recover from an unreviewed regression; team-scale collaborator roles.
- Branching trade-offs; Azure Repos policies and GitHub rulesets; Git LFS, Scalar, reflog, revert and cherry-pick; secret-history cleanup; the fixture's five delivery measures; AB# traceability, wikis, webhooks and collaboration tools. Current DORA definitions are reviewed separately.
Who you work with
Maya Chen
Senior AI Engineer · Your mentor
Kabelo Mokoena
Platform and DevOps Engineer
Leo Martins
Product Manager
Scenes
- 1.The rules nobody wrote down
- 2.What actually protects the team
- 3.Which branching strategy?
- 4.Read the DevOps dashboard
- 5.Make the repo enforce it
- 6.A normal fix, through the front door
- 7.Whose commit was that
- 8.Just this once
- 9.Contoso Creators: the delivery case study
- 10.Handover
What you leave with
You turned assist-service's team-scale rules into settings the repo enforces itself — a protected main, a required 'tests' check, and a promoted collaborator — shipped a normal fix through that policy, traced a display-name regression to the commit and author that caused it and recovered it properly attributed to you, and turned down a 'just this once' bypass request under real time pressure.