Signal / Noise

Applied GenAI Curriculum for AI PMs

How to make AI products

Over the past few years I've been organizing my learning and experience — shipping machine-learning, generative, and agentic systems in production — into a course. It is free, it asks nothing of you, and it exists to give back to the community. The CORE sequence (01–13) is a complete operating model on its own, capped by a capstone checkpoint; DEPTH units (14–24) go deeper on demand.

The capstone (Unit 13) is also the spec for the first artifact in ai-product-decisions — the course is the theory; the repo is the same material as working code.

24 units Free, no sign-up 24 of 24 published
Core Sequence · Units 01–13
01 What AI Engineering Is (and Isn't) Published
02 Why the Output Is Never the Same Twice Published
03 Should This Even Be an AI Feature? Published
04 Workflows vs. Agents: The Foundational Distinction Published
05 The Four-Way Decision: Prompting vs. RAG vs. Finetuning vs. Agents Published
06 Why Evals Are the Whole Game Published
07 Defining "Good": Evaluation Criteria & Model Selection Published
08 Prompting as an Engineering Discipline Published
09 Designing the Eval Pipeline Published
10 The Production Architecture, Assembled Published
11 Data Curation: The Unsexy Blocker Published
12 Cost & Latency: The Two Constraints That Kill Roadmaps Published
13 Capstone (CORE checkpoint) — interactive worksheet Checkpoint
Depth · Units 14–24 · pulled in on demand
14 Error Analysis & the Improvement Flywheel Published
15 LLM-as-Judge, Done Properly Published
16 Evals Reference Shelf Published
17 RAG Internals: Retrieval Quality Published
18 Agent Internals: Tools, Planning, Failure Modes, Memory Published
19 Finetuning: What It Actually Involves Published
20 Synthetic Data & Distillation Published
21 Monitoring, Observability & the Feedback Loop Published
22 Defensive Prompting: Injection, Jailbreaks, Extraction Published
23 Where Foundation Models Come From (optional background) Published
24 Foundation-Model Evaluation Concepts (optional background) Published
References & Further Reading Published
← Back home Questions or corrections are welcome — LinkedIn