How To Go From No Job Offer To Amazon's AGI Development Team In 18 Months - w/ Kunal

Here's the description, formatted for RSS/Spotify (timestamps on their own lines with spacing so they don't collapse): Seven years ago, Kunal Mishra was a summer intern at DRDO, facing the choice every engineering student knows: take the safe campus placement, or bet on the harder thing. He bet on machine learning. 🎯 That bet ran through a master's in the USA at Northeastern University and dropped him into the 2024 job market with no return offer and no sponsorship. Tough for anyone. Harder for an international student. So he built something. A 30-million-parameter LLM from scratch, trained for about $50 when the same build can cost thousands. (Yes, fifty dollars.) That one project got him his first US job. Then it got him Amazon. Today Kunal works on Amazon's AGI post-training team, building the Amazon Nova frontier models. This episode is the machine learning engineer roadmap he wishes someone had handed him. We open with what AGI actually means inside a big tech lab. Not the sci-fi version — the working definition his team uses to ship. Then the practical calls. Why fundamentals still beat chasing trends. How he chose between an MS in CS and an MS in AI (and why the label mattered less than he thought). What breaking into ML really looked like with no H1B sponsorship locked in. The heart of it is the project. Kunal walks through how to build an LLM from scratch — the design choices and the cheap-compute tricks that kept it under fifty bucks. A real llm project beats a polished resume every time. Then Amazon. The full interview loop, unfiltered. 450 LeetCode problems logged. Two DSA rounds. Two system design rounds — real ml system design, not whiteboard theater. And the 40% nobody preps enough for: the behavioral, mapped straight to Amazon's leadership principles. Plus how to hold your nerve through the bar raiser interview. He also hands over his prep shelf: Chip Huyen's Designing Machine Learning Systems and AI Engineering, Maxime Labonne's LLM Engineer's Handbook, Sebastian Raschka's blog and books, Andrej Karpathy's llama2.c, Hacker News (Show HN), and NeetCode 150. If you're a student or early-career engineer trying to break into ML in a 2026 market that feels stacked against you, this is your FAANG interview prep and your career advice in one sitting. Especially if you're doing it on hard mode as an international student. Press play. Then subscribe to Ready Set Do wherever you listen → readysetdopodcast.com New episodes on AI careers and the unconventional routes into big tech land every week. Chapters: 00:00 - Kunal's Journey: From Intern to ML Engineer 04:17 - Understanding AGI: Definitions and Perspectives 07:24 - The Importance of Specialization in Machine Learning 10:23 - Navigating Higher Education: Choosing the Right Program 13:16 - Job Market Insights: Finding Your First Role 16:27 - Building Projects: The Key to Job Success 19:10 - Lessons from Building an LLM from Scratch 22:15 - The Amazon Opportunity: Interview Process and Preparation 25:14 - System Design and Behavioral Interviews at Amazon 28:21 - Resources for Aspiring ML Engineers 31:28 - The Bar Raiser Experience at Amazon 34:12 - Working on Frontier Models at Amazon 37:15 - Final Thoughts: Health and Gratitude