Maven: Agentic AI Product Management Certification

Taught by Mahesh Yadav — 20-plus years in product roles at Google, Meta, AWS, and Microsoft — this cohort-based course on Maven targets product managers, engineers, and product leaders who want to build and ship AI-native products, particularly those that involve agents rather than simple LLM integrations. The next cohort runs May 9 through June 20, 2026, with over 21 live sessions and roughly six hours of engagement per week.

Who it is for

The course suits practitioners with some prior exposure to AI or ML concepts who want to go further than prompt engineering. It is well-matched to PMs at companies already building AI features who need to understand agent architectures, evaluation systems, and cost models. It is not a first introduction to product management — participants are expected to know how to write a PRD and run discovery interviews.

What it covers

The curriculum spans eight topic areas. It starts with ML fundamentals and agentic AI foundations, then moves into prompt engineering and AI product sense: how to identify which business problems actually benefit from AI, and which do not. A significant portion of the course addresses agent evaluation frameworks — deterministic and probabilistic testing, failure mode analysis, and token economics, all of which matter when building AI features for production at scale.

Students work directly with tools including Google AI Studio, Claude, n8n, OpenAI Agents SDK, and Azure AI services. The two major graded projects — a full PRD and a case study — are designed to serve as portfolio pieces in hiring conversations. Responsible AI practices run throughout: guardrails, hallucination handling, and model selection trade-offs are treated as product decisions, not engineering afterthoughts.

Weekly post-cohort sessions extend the learning after the six-week program ends, and all live sessions are recorded with lifetime access.

What it does not cover

The course does not address foundational product management skills or broad AI business strategy. It also does not teach coding — participants build working prototypes using no-code and low-code tools. At $3,000, it is positioned as a professional certification rather than a casual introduction. For PMs who want structured, expert-led training on agentic systems specifically, it is a strong candidate worth evaluating.