<![CDATA[AI Innovators: High School Edition]]>https://podcasters.spotify.com/pod/show/ai-innovators-high-schoolRiverside.fm (https://riverside.com)Sat, 17 Oct 2026 06:25:53 GMTMon, 06 Jul 2026 04:53:58 GMT60Sathvik SubramanianAI Innovators: High School Edition is a podcast that dives into the world of artificial intelligence through the voices of entrepreneurs, experts, and everyday professionals. In each episode, host Sathvik Subramanian explores how AI is transforming industries, empowering startups, and reshaping careers. From cutting-edge innovations to real-world applications, the podcast offers a deep look at the people and technologies driving the AI revolution. Tune in for insightful conversations and a collective exploration of AI’s impact on the future.episodicSathvik Subramanianinfoglean.nyc@gmail.comno<![CDATA[The first generation not to get smarter. Dr. Tracey Tokuhama-Espinosa (Harvard): Learning Vs AI.]]>This is the third episode in a series about how humans learn and how machines learn. Sumitra Ganesh (J.P. Morgan) told me small models are the most underhyped idea in AI because the brain runs on twenty watts. Anish Mahishi (NVIDIA) told me reasoning models finally work because they think one internal token at a time, the way humans do. Both answers pointed at the human brain. Dr. Tracey Tokuhama-Espinosa from Harvard has spent her career studying it.

Dr. Tracey Tokuhama-Espinosa teaches the Neuroscience of Learning at Harvard and has spent three decades studying how humans actually learn. In this conversation she explains neuroconstructivism, the Five Pillars of the Mind, why cognitive offloading via ChatGPT literally removes the scaffolding your brain needs for higher-order thinking, why yours is the first generation not to get smarter, and the one non-technical piece of advice she gave a fifteen-year-old host.

Guest

]]>c3b43e00-e28f-4c1c-9a67-e4d04f253babMon, 31 Aug 2026 23:42:26 GMT<p>This is the third episode in a series about how humans learn and how machines learn. Sumitra Ganesh (J.P. Morgan) told me small models are the most underhyped idea in AI because the brain runs on twenty watts. Anish Mahishi (NVIDIA) told me reasoning models finally work because they think one internal token at a time, the way humans do. Both answers pointed at the human brain. Dr. Tracey Tokuhama-Espinosa from Harvard has spent her career studying it.<br /><br />Dr. Tracey Tokuhama-Espinosa teaches the Neuroscience of Learning at Harvard and has spent three decades studying how humans actually learn. In this conversation she explains neuroconstructivism, the Five Pillars of the Mind, why cognitive offloading via ChatGPT literally removes the scaffolding your brain needs for higher-order thinking, why yours is the first generation not to get smarter, and the one non-technical piece of advice she gave a fifteen-year-old host.<br /><br /><b>Guest</b></p><ul><li><a rel="noopener noreferrer nofollow" href="https://scholar.google.com/citations?user=wGU0gj4AAAAJ" target="_blank">Dr. Tracey Tokuhama-Espinosa. Google Scholar</a></li><li><a rel="noopener noreferrer nofollow" href="https://www.linkedin.com/in/tracey-tokuhama-espinosa-a8532460/" target="_blank">LinkedIn</a></li><li><a rel="noopener noreferrer nofollow" href="https://coursebrowser.dce.harvard.edu/course/neuroscience-of-learning-an-introduction-to-mind-brain-health-and-education/" target="_blank">Neuroscience of Learning course at Harvard Extension</a></li><li><a rel="noopener noreferrer nofollow" href="https://thelearningsciences.com/" target="_blank">Conexiones: The Learning Sciences Platform</a><br /><br /><b>Books mentioned or referenced</b><ul><li><a rel="noopener noreferrer nofollow" href="https://wwnorton.com/books/9780393713213" target="_blank">Five Pillars of the Mind: Redesigning Education to Suit the Brain (W. W. Norton, 2019)</a></li><li><a rel="noopener noreferrer nofollow" href="https://wwnorton.com/books/9780393706079" target="_blank">Mind, Brain, and Education Science (W. W. Norton, 2010)</a></li><li><a rel="noopener noreferrer nofollow" href="https://www.tcpress.com/bringing-the-neuroscience-of-learning-to-online-teaching-9780807765524" target="_blank">Bringing the Neuroscience of Learning to Online Teaching (Teachers College Press)</a></li></ul><p><b>Concepts and research referenced in the conversation</b></p><ul><li><a rel="noopener noreferrer nofollow" href="https://en.wikipedia.org/wiki/Flynn_effect" target="_blank">Flynn effect (the rise, and recent decline, of average IQ)</a></li><li><a rel="noopener noreferrer nofollow" href="https://www.hbs.edu/ris/Publication%20Files/11-091.pdf" target="_blank">The IKEA effect (Norton, Mochon, Ariely)</a></li><li><a rel="noopener noreferrer nofollow" href="https://en.wikipedia.org/wiki/Neuroconstructivism" target="_blank">Neuroconstructivism (Karmiloff-Smith / Sirois / Westermann)</a></li><li><a rel="noopener noreferrer nofollow" href="https://abcdstudy.org/" target="_blank">Adolescent Brain Cognitive Development (ABCD) Study</a></li><li><a rel="noopener noreferrer nofollow" href="https://www.adultdevelopmentstudy.org/" target="_blank">The Harvard Study of Adult Development (nearly a century long)</a></li></ul><p><b>Host</b></p><ul><li><a rel="noopener noreferrer nofollow" href="https://www.wisenai.com/" target="_blank">Wisen AI. Sathvik's startup, creating training data by paying senior citizens for their tacit knowledge</a></li><li><a rel="noopener noreferrer nofollow" href="https://www.newarka.edu/" target="_blank">Newark Academy</a></li></ul><p><b>Contact</b> If you're a high schooler building something with AI, especially at the edge of AI and learning, send it to Sathvik: <a rel="noopener noreferrer nofollow" href="mailto:sathvik@wisenai.com" target="_blank">sathvik@wisenai.com</a></p></li></ul><p></p>no01:11:07The first generation not to get smarter. Dr. Tracey Tokuhama-Espinosa (Harvard): Learning Vs AI.full<![CDATA[The AI you use was built twice - Anish Mahishi (NVIDIA NeMo-RL) on post-training]]>Anish Mahishi is an engineer on the NeMo-RL team at NVIDIA. He works on the post-training systems that turn raw pretrained models into the AI you actually use — ChatGPT, Claude, and every real product built on top of a language model. Before NVIDIA he was a data scientist at Fractal, a Master's student at NYU (with medical MRI research at NYU Langone), an intern on the PyTorch team at Meta, and a machine learning engineer at Bloomberg for three years.

In this episode we cover:

  • How he got into AI (an Andrew Ng Coursera course in his second year of college)
  • The full pipeline from you hitting enter in ChatGPT to the first word showing up
  • What post-training actually is, and what the NeMo-RL team builds
  • RLHF vs. DPO vs. GRPO — the alphabet soup, translated
  • What DeepSeek R1 actually changed about how these systems run
  • Why fully autonomous agents are overhyped and evaluations are underhyped
  • Why medical AI is the hardest domain to get right
  • On-device AI, and the one non-technical skill that mattered more than he expected

Guest

Anish's research

Concepts and tools we discussed

Contact If you're a high schooler building something with AI, send it to Sathvik: sathvik@wisenai.com

]]>8ba7acde-e123-4240-bf34-7b1765c04576Mon, 27 Jul 2026 03:15:35 GMT<p>Anish Mahishi is an engineer on the NeMo-RL team at NVIDIA. He works on the post-training systems that turn raw pretrained models into the AI you actually use — ChatGPT, Claude, and every real product built on top of a language model. Before NVIDIA he was a data scientist at Fractal, a Master's student at NYU (with medical MRI research at NYU Langone), an intern on the PyTorch team at Meta, and a machine learning engineer at Bloomberg for three years.</p><p>In this episode we cover:</p><ul><li>How he got into AI (an Andrew Ng Coursera course in his second year of college)</li><li>The full pipeline from you hitting enter in ChatGPT to the first word showing up</li><li>What post-training actually is, and what the NeMo-RL team builds</li><li>RLHF vs. DPO vs. GRPO — the alphabet soup, translated</li><li>What DeepSeek R1 actually changed about how these systems run</li><li>Why fully autonomous agents are overhyped and evaluations are underhyped</li><li>Why medical AI is the hardest domain to get right</li><li>On-device AI, and the one non-technical skill that mattered more than he expected</li></ul><p><b>Guest</b></p><ul><li><a rel="noopener noreferrer nofollow" href="https://scholar.google.com/citations?user=zu_hErcAAAAJ" target="_blank">Anish Mahishi — Google Scholar</a></li><li><a rel="noopener noreferrer nofollow" href="https://www.linkedin.com/in/anishmahishi/" target="_blank">Anish Mahishi — LinkedIn</a></li></ul><p><b>Anish's research</b></p><ul><li><a rel="noopener noreferrer nofollow" href="https://arxiv.org/abs/2301.11962" target="_blank">On the feasibility of machine learning augmented magnetic resonance for point-of-care identification of disease (arXiv 2301.11962, 2023)</a> </li><li><a rel="noopener noreferrer nofollow" href="https://arxiv.org/abs/2605.01417" target="_blank">Medmarks: A Comprehensive Open-Source LLM Benchmark Suite for Medical Tasks (arXiv 2605.01417, May 2026)</a> </li><li><a rel="noopener noreferrer nofollow" href="https://github.com/MedARC-AI/Medmarks" target="_blank">Medmarks on GitHub</a></li></ul><p><b>Concepts and tools we discussed</b></p><ul><li><a rel="noopener noreferrer nofollow" href="https://www.coursera.org/learn/machine-learning" target="_blank">Andrew Ng's Machine Learning course on Coursera</a> </li><li><a rel="noopener noreferrer nofollow" href="https://github.com/NVIDIA/NeMo-RL" target="_blank">NVIDIA NeMo-RL on GitHub</a> </li><li><a rel="noopener noreferrer nofollow" href="https://arxiv.org/abs/2203.02155" target="_blank">Training language models to follow instructions with human feedback (InstructGPT / RLHF paper)</a></li><li><a rel="noopener noreferrer nofollow" href="https://arxiv.org/abs/2305.18290" target="_blank">Direct Preference Optimization (DPO paper)</a></li><li><a rel="noopener noreferrer nofollow" href="https://arxiv.org/abs/2402.03300" target="_blank">DeepSeekMath — GRPO paper</a></li><li><a rel="noopener noreferrer nofollow" href="https://arxiv.org/abs/2501.12948" target="_blank">DeepSeek-R1 — the reasoning-model recipe we discussed</a></li><li><a rel="noopener noreferrer nofollow" href="https://pytorch.org/" target="_blank">PyTorch</a> </li></ul><ul><li><a rel="noopener noreferrer nofollow" href="https://www.wisenai.com/" target="_blank">Wisen AI</a> — Sathvik's startup, bridging older-adult experience and modern AI</li><li><a rel="noopener noreferrer nofollow" href="https://www.newarka.edu/" target="_blank">Newark Academy</a></li></ul><p><b>Contact</b> If you're a high schooler building something with AI, send it to Sathvik: <a rel="noopener noreferrer nofollow" href="mailto:sathvik@wisenai.com" target="_blank">sathvik@wisenai.com</a></p>no00:42:07The AI you use was built twice - Anish Mahishi (NVIDIA NeMo-RL) on post-trainingfull<![CDATA[AI Agents - Hype vs Real Research. A Conversation with Dr. Sumitra Ganesh of JP Morgan]]>Dr. Sumitra Ganesh heads AI Research in JP Morgan. Enjoy a fascinating conversation about her research.

1. The 2019 dealer-market paper

2. Phantom

3. The 2021 paper

4. The 2020 NeurIPS paper Dr. Sumitra referenced

5. The 2022 paper that extended the framework to OTC markets

6. The 2025 Annual Reviews piece

The paper she said most influenced her thinking recently

The tool she said surprised her most in the last two years - Claude Code

WIsenAI.com - Sathvik's start up. He reads everything - send your comments to sathvik@wisenai.com

]]>8c763af5-bd27-4a50-8cf6-2de7b22fc6efMon, 06 Jul 2026 05:02:11 GMT<p>Dr. Sumitra Ganesh heads AI Research in JP Morgan. Enjoy a fascinating conversation about her research.</p><ul><li><a rel="noopener noreferrer nofollow" href="https://www.jpmorgan.com/technology/artificial-intelligence/about/sumitra-ganesh" target="_blank">⁠Dr. Sumitra Ganesh — bio page at J.P. Morgan AI Research⁠</a></li><li><a rel="noopener noreferrer nofollow" href="https://scholar.google.com/citations?user=B6t9KM8AAAAJ" target="_blank">⁠Sumitra Ganesh on Google Scholar⁠</a> — full publication list</li><li><a rel="noopener noreferrer nofollow" href="https://www.jpmorgan.com/technology/artificial-intelligence/research" target="_blank">⁠J.P. Morgan AI Research⁠</a></li></ul><p><b>1. The 2019 dealer-market paper</b></p><ul><li><a rel="noopener noreferrer nofollow" href="https://arxiv.org/abs/1911.05892" target="_blank">⁠Reinforcement Learning for Market Making in a Multi-agent Dealer Market (Ganesh et al., NeurIPS 2019)⁠</a></li></ul><p><b>2. Phantom</b></p><ul><li><a rel="noopener noreferrer nofollow" href="https://github.com/jpmorganchase/Phantom" target="_blank">⁠Phantom on GitHub⁠</a></li><li><a rel="noopener noreferrer nofollow" href="https://arxiv.org/abs/2210.06012" target="_blank">⁠Phantom — A RL-driven Multi-Agent Framework to Model Complex Systems (Ardon, Vann, Garg, Spooner, Ganesh — AAMAS 2023)⁠</a></li><li><a rel="noopener noreferrer nofollow" href="https://phantom-marl.readthedocs.io/" target="_blank">⁠Phantom documentation⁠</a></li></ul><p><b>3. The 2021 paper</b></p><ul><li><a rel="noopener noreferrer nofollow" href="https://arxiv.org/abs/2110.06829" target="_blank">⁠Towards a fully RL-Based Market Simulator (Ardon, Vadori, Spooner, Xu, Vann, Ganesh — ICAIF 2021)⁠</a></li></ul><p><b>4. The 2020 NeurIPS paper Dr. Sumitra referenced</b></p><ul><li><a rel="noopener noreferrer nofollow" href="https://proceedings.neurips.cc/paper/2020/hash/a2f04745390fd6897d09772b2cd1f581-Abstract.html" target="_blank">⁠Calibration of Shared Equilibria in General Sum Partially Observable Markov Games (Vadori, Ganesh, Reddy, Veloso — NeurIPS 2020)⁠</a></li></ul><p><b>5. The 2022 paper that extended the framework to OTC markets</b></p><ul><li><a rel="noopener noreferrer nofollow" href="https://arxiv.org/abs/2210.07184" target="_blank">⁠Towards Multi-Agent Reinforcement Learning driven Over-The-Counter Market Simulations (Vadori, Ardon, Ganesh et al. — 2022; Mathematical Finance, 2024)⁠</a></li></ul><p><b>6. The 2025 Annual Reviews piece</b></p><ul><li><a rel="noopener noreferrer nofollow" href="https://www.annualreviews.org/content/journals/10.1146/annurev-control-022823-033813" target="_blank">⁠Generative AI Agents for Knowledge Work Augmentation in Finance (Annual Reviews of Control, Robotics, and Autonomous Systems, Vol. 8, May 2025)⁠</a></li></ul><p><b>The paper she said most influenced her thinking recently</b></p><ul><li><a rel="noopener noreferrer nofollow" href="https://news.mit.edu/2022/artificial-intelligence-can-detect-parkinsons-from-breathing-patterns-0822" target="_blank">⁠Artificial intelligence model can detect Parkinson's from breathing patterns (MIT News, Aug 2022)⁠</a></li><li><a rel="noopener noreferrer nofollow" href="https://www.eecs.mit.edu/people/dina-katabi/" target="_blank">⁠Dina Katabi — MIT EECS faculty page⁠</a></li></ul><p><b>The tool she said surprised her most in the last two years - Claude Code</b><br /><br /><a rel="noopener noreferrer nofollow" href="http://WIsenAI.com" target="_blank">WIsenAI.com</a> - Sathvik's start up. He reads everything - send your comments to <a rel="noopener noreferrer nofollow" href="mailto:sathvik@wisenai.com" target="_blank"><b>sathvik@wisenai.com</b></a></p>no00:52:53AI Agents - Hype vs Real Research. A Conversation with Dr. Sumitra Ganesh of JP Morganfull<![CDATA[Welcome!]]>In this brief introduction episode, host Sathvik Subramanian kicks off AI Innovators: High School Edition by giving you a quick overview of what the podcast is all about. Tune in to learn how we'll be exploring the exciting world of artificial intelligence, featuring insights from entrepreneurs, experts, and everyday professionals. Whether you're curious about AI’s impact on industries, careers, or startups, this podcast is your gateway to understanding the future of technology and anything AI-related. Get ready for conversations that inspire, inform, and spark your curiosity about the AI revolution!

]]>https://podcasters.spotify.com/pod/show/ai-innovators-high-school/episodes/Welcome-e2t2mqo00be8a8f-9f0a-4435-9cf8-b8d05efed821Sun, 05 Jan 2025 00:17:17 GMT<p>In this brief introduction episode, host Sathvik Subramanian kicks off <em>AI Innovators: High School Edition</em> by giving you a quick overview of what the podcast is all about. Tune in to learn how we'll be exploring the exciting world of artificial intelligence, featuring insights from entrepreneurs, experts, and everyday professionals. Whether you're curious about AI’s impact on industries, careers, or startups, this podcast is your gateway to understanding the future of technology and anything AI-related. Get ready for conversations that inspire, inform, and spark your curiosity about the AI revolution!</p> no00:01:08Welcome!full