Prompt engineering is now a core AI PM skill. The right course makes you fluent in weeks. Ranked from quick foundations to deep technique.
Why Prompt Engineering Is a Core AI PM Skill
The "is prompt engineering dead" debate is over. The answer is no — it just looks different than it did in 2023. Modern prompt engineering is part product spec, part system design, part eval discipline. AI PMs who are fluent in prompting ship faster, reason more clearly about model behavior, and write better engineering specs. AI PMs who aren't fluent end up dependent on engineers for the most important product decisions.
You don't need to become a prompt engineer. You need to be comfortable enough that you can sketch prompts in product specs, debug failures live with engineers, and run your own quick experiments without filing a ticket. The courses below get you there. They are ranked by what they're best at — foundations, technique depth, multimodal, or developer integration — not by overall "quality," because the right course depends on what you need to learn next.
🎓Want hands-on AI PM training that goes beyond prompting? The AI PM Masterclass covers prompt engineering, evals, and production AI in 4 weekends.
Best Quick Foundations
1. ChatGPT Prompt Engineering for Developers (DeepLearning.AI)
Andrew Ng and Isa Fulford's short course remains the single best place to start. It's roughly an hour of focused video, with hands-on Jupyter notebooks. The curriculum covers principles (specificity, structured output, iterative refinement) and immediately applies them to summarization, transformation, and inference tasks.
The course is free, fast, and gets you to "I can write a useful prompt" in an afternoon. It's not the deepest course on this list, but it's the best ROI for a working PM.
Why AI PMs need this: The fastest path from zero to functional. Block out one afternoon and finish it.
Visit Course2. Anthropic Prompt Engineering Interactive Tutorial
Anthropic's free interactive tutorial is the best vendor-published prompting curriculum in 2026. It runs as a hands-on Jupyter notebook (or hosted in their console) and walks you through structuring prompts for Claude with concrete exercises. The "what works, what doesn't" examples are especially good — you watch identical tasks succeed and fail based on prompt structure.
Even if you don't ship on Claude, the principles (XML tagging, role specification, structured reasoning, output constraints) generalize directly to GPT and Gemini.
Why AI PMs need this: The clearest, most concrete prompting tutorial available. Especially valuable if your team uses Claude.
Visit Tutorial3. Generative AI for Everyone (Andrew Ng, Coursera)
Andrew Ng's broader generative AI course is the right starting point if you need conceptual grounding before tactical prompting. It covers what LLMs are, what they can and can't do, common product patterns, and the role of prompting within the larger stack.
For PMs new to AI, this is the most accessible entry point on the list. Pair it with the DeepLearning.AI short course and you have a complete two-day foundation.
Why AI PMs need this: Best big-picture course for PMs new to AI. Light on technique but strong on context. Pair with our beginner's guide.
Visit CourseBest Comprehensive Curricula
4. Prompt Engineering Specialization (Vanderbilt, Coursera)
Vanderbilt's three-course specialization led by Jules White is the deepest academic treatment of prompt engineering currently available on a major platform. It introduces named prompt patterns (persona, audience, template, recipe, output automater) and walks through them with multiple examples. The "Prompt Engineering for ChatGPT" first course alone is excellent.
The course occasionally over-formalizes (you don't strictly need named patterns to be effective), but the rigor makes it especially valuable for PMs who want a shared vocabulary with their engineering team.
Why AI PMs need this: The most thorough structured curriculum. Best for PMs who learn through frameworks and shared vocabulary.
Visit Course5. Microsoft AI Engineer Learning Path
Microsoft's official Generative AI for Beginners curriculum (and the broader AI Engineer learning path on Microsoft Learn) is the strongest free integrated curriculum from a major vendor. It covers prompting, Azure OpenAI patterns, RAG, fine-tuning, and responsible AI in modular bites you can take at your own pace.
Particularly strong if you're shipping into Microsoft's ecosystem (Azure, Copilot Studio, Power Platform), but the prompting fundamentals are vendor-neutral.
Why AI PMs need this: The most comprehensive free curriculum from a major vendor. Excellent if your stack touches Azure.
Visit CourseBest Open References and Guides
6. Promptingguide.ai (DAIR.AI)
The DAIR.AI Prompt Engineering Guide is the most respected open reference site. It covers every major prompting technique — zero-shot, few-shot, chain-of-thought, ReAct, self-consistency, Tree of Thoughts, and more — with citations to the underlying research. It's updated frequently and is the canonical reference for technique terminology.
Not a course in the traditional sense, but a working PM should be able to navigate it fluently. Read it once end-to-end, then keep it as a reference. See our chain-of-thought primer.
Why AI PMs need this: The technical reference. Read once, bookmark forever.
Visit Guide7. LearnPrompting.org
LearnPrompting is a long-running open curriculum that covers prompting from beginner to advanced, including a strong red-teaming and prompt-security section that few other courses address. The hosted "HackAPrompt" challenges teach adversarial thinking better than any other resource — invaluable for PMs working on agent or enterprise products.
The pacing varies (some chapters feel rushed, others are deep), but as a free resource it is exceptional.
Why AI PMs need this: The best free resource for prompt security and adversarial thinking. Essential for agent and enterprise PMs.
Visit Guide8. OpenAI Cookbook
The OpenAI Cookbook is a vast repository of working code examples, prompting recipes, and integration patterns. It's not a sequential course, but if you learn by reading working code and adapting it, the Cookbook will teach you more than most paid courses.
Particularly strong for structured output, function calling, agents, and embeddings. AI PMs who spend a few hours skimming it learn the texture of what's possible.
Why AI PMs need this: The recipe book. Read by example to learn what "good" looks like across many use cases.
Visit CookbookBest Cohort and Live Courses
9. Maven Cohort Courses (AI PM and Eval-Focused)
Maven hosts the strongest collection of cohort-based AI product courses in 2026 — including Aman Khan's AI PM course, Hugo Bowne-Anderson's LLM evaluation course, and several specialized prompting workshops from working practitioners. The format is intense, expensive, and high-signal: live sessions, real peer cohorts, and instructors who are actively shipping.
If async self-paced courses leave you adrift, cohort courses are the antidote. You'll spend more, finish more, and build a peer network in the process.
Why AI PMs need this: The best paid cohort option for PMs who finish what they start when there's structure and peer pressure.
Visit Maven10. Building Systems with the ChatGPT API (DeepLearning.AI)
The follow-up to "Prompt Engineering for Developers" focuses on chaining prompts together to build end-to-end systems. It's the right next step for PMs who finished the first short course and want to see how prompting composes into real applications — input classification, moderation, chained reasoning, evaluation.
Especially useful for AI PMs writing PRDs that involve multi-step LLM workflows. After this course you'll write much more realistic spec diagrams.
Why AI PMs need this: Bridges prompting and system design. Levels you up from "single prompt" to "real application."
Visit CourseLearning Strategy
Pick one short course (Andrew Ng or Anthropic) and finish it this week. Spend the next two weeks practicing on a real product problem at work. Then layer in one comprehensive course or a cohort. Most PMs fail by enrolling in five courses and finishing none. The fastest learners finish one, apply it, then pick the next gap to fill.
How to Choose Your First Course
Pick based on your current gap, not on credentials.
New to AI? Start with "Generative AI for Everyone." Build conceptual context, then layer in tactical prompting.
Need to ship this quarter? Do the DeepLearning.AI short course and the Anthropic tutorial in the same week. You'll be functional immediately.
Want depth? Vanderbilt's specialization plus the DAIR.AI guide will give you the strongest conceptual vocabulary.
Working on safety or agents? LearnPrompting.org's security chapters are the most useful resource available.
What Courses Won't Teach You
The best courses give you techniques. They cannot give you taste. Taste comes from practice — running real prompts against real evals on real product problems. Plan to spend at least two hours practicing for every hour of video you consume. Anything less and the courses become entertainment.
The other thing courses won't teach you is when not to use prompting. Sometimes the right answer is fine-tuning, sometimes it's a deterministic system, sometimes it's not using AI at all. Senior AI PMs make these calls regularly. Pair your prompting practice with broader product thinking. See our prompt engineering guide for context.
Building a Daily Practice
Courses get you started. Practice keeps you sharp. Three habits separate fluent AI PMs from rusty ones.
One prompt experiment per week. Pick a real problem (from work or a side project). Try three prompt variations. Compare outputs. Write down what you learned.
Read one model card per month. When a new model ships, read the system card. You'll quickly internalize how model design affects product behavior.
Keep a prompt journal. Save the prompts that worked, the ones that failed, and what changed. After three months, this becomes the most valuable resource you have.
Your Prompting Curriculum
You don't need every course on this list. You need to finish two, practice for a month, and then pick the next gap. Skip the credential collecting. Optimize for fluency.
Want structured cohort training that goes beyond prompting into evals, agents, and shipping? Our AI Product Management Masterclass integrates prompting practice into a full AI PM curriculum, with senior instructors and a working peer cohort.