Building with the Claude API.

New Week, New learning 🎓.

Anthropic recently launched its Academy, 13 free self-paced high quality courses on Skilljar, covering prompt engineering, building with the Claude API, and agentic workflows.

For anyone working with LLMs in product, GTM, or applied research, this is worth a look. The catalog covers real craft questions, how to structure prompts for specific outputs, when to use tool use vs. plain prompting, how agentic workflows change what's possible, rather than generic "AI for beginners" content.

I've started with the API course and I'm testing it against a craft question: e.g. structuring the research-to-narrative pipeline - how you move from dense primary material (interviews, scientific papers, technical documentation, competitive signals) to a clear set of positioning hypotheses and narrative frames. 

Choosing what to automate is the part that's most interesting to me and the part Anthropic's materials actually address, which distinguishes their courses from a lot of what's out there.

I'll share my learning over the next few weeks, particularly around how this kind of workflow changes positioning work, research synthesis, and GTM content production for complex or regulated offers. More soon!

Eval workflow

System prompts