AI agent PM roles are the fastest-growing, highest-paid positions in tech. These jobs pay $750K+ (TC in SF/NY) and are growing 2-3x faster than traditional PM roles. But most people don't know how to actually build AI agents. They think it's just ChatGPT with a fancy interface. Today I sat down with Mahesh Yadav, who's worked as a PM at Meta, Amazon, Microsoft, and Google. He's built AI agents at scale for 8+ years and now teaches hundreds of PMs at top companies. He breaks down the exact playbook: how to build agents, the 18-month roadmap to $750K+ roles, and what FAANG companies look for in vibe coding interviews. If you want to learn to build AI agents , this is your blueprint . Check out the conversation on Apple , Spotify and YouTube . ---- Brought to you by: * Maven: Get $100 off my curation of their top courses with code ‘AAKASH550C7’ * Miro : The innovation workspace is your team’s new canvas * Kameleoon : Leading AI experimentation platform * The AI Evals Course for PMs & Engineers : Get $1155 off with code ‘ag-evals’ * Amplitude : The market-leader in product analytics ---- Timestamps 00:00 - Introduction & Overview 01:40 - What Makes an AI Agent PM 02:37 - Building the Backend Agent 16:32 - Creating the Frontend with V0 25:27 - What Defines an AI Agent vs AI Product 30:15 - AI PM Interview Requirements 34:08 - Cart Before the Horse Development 37:15 - Breaking into FAANG: Mahesh's Story 42:17 - Internal Transfer Strategy 50:40 - Comparing Microsoft vs Amazon vs Meta vs Google 54:28 - AI Agent PM Job Market & Salary Data 57:26 - Can Anyone Become an AI PM? 59:14 - 18-Month Roadmap to AI PM 1:05:01 - AI Agents for Regular PMs 1:08:47 - Business of Mahesh & Course Success ---- 10 Steps to a $750K+ AI Agents Job: 1. Build First (Not Study) The biggest mistake aspiring AI PMs make is spending months reading about AI instead of building. Companies like Google aren't looking for people who know frameworks—they want builders who have actually shipped AI products. Start with tools like Langflow for no-code backends and V0 for frontends. 2. Master AI Fundamentals You need to know how models work, how data contributes to these models, and how to evaluate agent performance. Can you make smart choices between different models? Do you understand how these models are built and how to interact with them? This knowledge separates real AI PMs from pretenders. 3. Show Scale Experience FAANG companies desperately need people who have seen one major technology transition and navigated it successfully. Whether it was cloud migration, mobile, or something else, show you can handle the chaos that comes with emerging tech. They're looking for people who experiment constantly because AI is new for everyone. 4. Prototype in Weeks The cost of prototyping has dropped 100x in two years. Instead of spending six months on research and PRDs, build a working prototype in 2-3 weeks and show it to customers. This "cart before the horse" approach is now the competitive advantage in AI product development. 5. Get 10-20 Real Users Find a real problem you can solve—ideally one where you have PhD-level expertise, involves unstructured data, and requires complex decision-making. Build an agent to solve it and get at least 10-20 people actually using it. This teaches you evaluation and iteration in ways no course can. 6. Scale to Production Hire a small team of engineers (even remotely) and get your prototype into real production with 100+ users. This teaches you the difference between a demo and a scalable system. Many startups will let you do this for free in exchange for the experience and expertise you bring. 7. Target Dream Companies Pick your top 10 target companies and start contributing to their open communities. Run evaluations on their products for free. Show them gaps in their AI capabilities. Build features for their open-source models. Make yourself impossible to ignore by doing the work their PMs should be doing. 8. Master Vibe Coding In vibe coding interviews, they're not testing your technical skills—they're judging your product thinking. Show structured prompts, demonstrate how you iterate based on user feedback, and prove you can evaluate and improve AI systems. Practice the three-step framework: task, requirements, resources. 9. Negotiate Multiple Offers AI PM roles at FAANG companies pay $750K-$1.5M+ total comp because demand far exceeds supply. Don't settle for one offer. The best candidates often get rejected by one company only to get double the salary elsewhere. Persistence pays—literally. 10. Execute 18-Month Timeline Month 1-3: Learn fundamentals and build your first agent. Month 4-6: Get 10-20 real users on a product you built. Month 7-12: Scale to production with 100+ users. Month 13-18: Contribute to target companies and interview. This timeline works because there's a level playing field in AI—your background matters less than your ability to ship. ---- Related Podcasts: * AI Agents for PMs in 69 Minutes * Fu
