Part one · On YouTube
Fundamentals
How generative AI actually works, why context changes the answer, and the four parts of a prompt — persona, instruction, context, format.
Free course — two parts
Let's Make Prompts Work
AI answers are inconsistent in a specific way: sometimes exactly right, sometimes vague, sometimes confidently wrong. The model didn't change between those three — your prompt did. This is a short course on writing the kind that works.
One course, two parts, both free. Part one is on YouTube. Part two arrives by email.
Start here
Fundamentals of generative AI and prompt engineering
What the model is actually doing when it answers, why context changes the result so much, and the four parts of a prompt — persona, instruction, context, format.
Get part two
One course, two parts, both free. Part one is on YouTube. Part two arrives by email.
- Part two, straight away
- The complete slide deck, as a PDF
Both parts are free, and that is the whole course — there is no part three waiting behind a price. This is the first course in a longer Prompt Engineering path; the courses after it go deeper and those will be paid. I'd rather you know that now than find out from an email.
The two parts
Part two · By email
Judging it, then using it
Evaluating and refining what comes back — accuracy, coherence, bias, and what to change first when a response almost works. Then putting it to work on real tasks.
Where this goes, if you want more
Four tracks, each going considerably deeper than this course does. None of them are out yet, and they won't be free — but this one stays free permanently.
Prompt Engineering
Beyond the fundamentals: evaluation, structured output, and prompts that survive being used a thousand times.
Model Context Protocol
Giving a model real tools and real context, not just better wording.
Agentic AI for developers
Systems that plan and act, and the parts that actually break in production.
AI-assisted coding
Working with an agent on a real codebase without losing the thread.