Foundation · Available
Think Like an AI Engineer
TAI-TAEThe engineering mental model: what a model is doing, how to choose one, how to give it context, and how to catch the failure a demo would hide.
- Effort
- 6 weeks
- Level
- Professional
- Passing
- 70%
Competency blueprint
Assessment weighting is published up front — 20% knowledge, 30% labs, 20% portfolio and 30% capstone.
Knowledge checks
20%A quiz per module, marked immediately.
Practical labs
30%Work done in the tools, submitted as evidence.
Portfolio
20%The assignments that accumulate across the track.
Capstone
30%A realistic brief, graded against a rubric by a human.
Assessment blueprint
Knowledge
20%
Labs
30%
Portfolio
20%
Capstone
30%
70% overall to pass. The capstone must pass independently of every other score. Two retakes per assessment window.
Capstone
Build something small that uses a model in its core loop — a script, a workflow, a tool — with an eval set of at least five cases including one it fails. Submit the code, the eval results, and what you'd change.
Modules
- How Modern AI Actually Works
- Prompting Like a Pro
- Picking the Right Model for the Job
- Context Is Everything
- Working With Files, Images, and Your Own Data
- AI for Coding
- Tool Use and Agents
- Building Simple AI Workflows Without Code
- Evals
- Staying Current Without Drowning