Show notes
Ever wondered how Amazon builds generative AI for millions of sellers? Or how Instagram's recommendation feed knows exactly what you want to watch next? In this deep-dive conversation, we sit down with AI/ML leaders from Amazon and Meta to uncover the real strategies behind building AI products at scale.Anita shares her journey launching Amazon's first AI image generation solution for sellers, while our Meta engineer breaks down distributed inference and how Instagram's recommendation models actually work.Key Insights:ā Why you should validate AI ideas with free tools (Midjourney, Canva) BEFORE buildingā The real difference between AI metrics vs. business metricsā How to define "quality" when there's no industry benchmarkā Why giving users full control over prompts is a mistakeā How Instagram updates its models without losing your preferencesTimeline:0:00 - Introduction: Building AI products at scale1:19 - Launching Amazon's first AI image generation tool2:06 - Balancing innovation with customer problems3:16 - The problem: Small sellers can't afford graphic designers4:03 - Real-world example: Food tech & restaurant images5:07 - Validating AI with prototypes before building5:24 - KEY INSIGHT: Use Midjourney/Canva to validate first6:12 - Quality dimensions: Aesthetics, relevance, proportions8:48 - Product manager's dilemma: AI metrics vs. business metrics9:30 - Creating benchmarks when none exist10:30 - Responsible AI: Safety, watermarks, artifacts11:04 - Business metrics: Adoption, engagement, retention12:05 - Defining accuracy in generative AI13:56 - Don't make users prompt engineers (abstract the complexity)15:26 - Fundamentals of inference explained16:09 - Training vs. Inference: The dog analogy17:00 - Why training and inference aren't binary18:43 - How Meta does distributed inference19:32 - How Instagram recommendations actually work20:26 - Snapshot updates: Keeping models fresh21:01 - Replacing models without losing user context22:38 - What is distributed inference? Tree structure explained23:31 - How Instagram serves personalized content at scaleWho This Is For:Product managers building AI/ML productsEngineers working on generative AIStartup founders exploring AI solutionsAnyone curious about how Big Tech AI actually worksResources Mentioned:Stable Diffusion aesthetic modelsMidjourney for prototypingCanva for quick validationš Subscribe for more deep dives into AI product development!#GenerativeAI #MachineLearning #ProductManagement #Amazon #Meta #Instagram
