{"type":"rich","version":"1.0","provider_name":"Transistor","provider_url":"https://transistor.fm","author_name":"Retailgentic | Consumer Behavior & Retail Trend","title":"From Columbia to Custom AI, an Agentic Commerce Deep Dive | Amine Allouah, My Custom AI","html":"<iframe width=\"100%\" height=\"180\" frameborder=\"no\" scrolling=\"no\" seamless src=\"https://share.transistor.fm/e/31664335\"></iframe>","width":"100%","height":180,"duration":3739,"description":"Dr. Amine Allouah completed his PhD at Columbia Business School, specializing in algorithmic game theory and optimization in multi-agent systems. After leading applied science work at Meta across ads, notifications, and marketplace teams, he co-founded My Custom AI to help enterprises deploy tailored AI solutions. His work bridges deep theory with practical applications, shaping how businesses think about LLMs, agents, and the economics of AI.\nThis week on Retailgentic, we’re joined by Dr. Amine, co-founder of My Custom AI and one of the sharpest minds at the intersection of AI, economics, and retail. In this conversation, we dig deep into:\nAmine’s journey: École Polytechnique → Columbia → Meta → entrepreneurship\nWhat algorithmic game theory means for real-world AI systems\nWhy enterprises need custom models for accuracy, privacy, and cost savings\nThe ACES Framework: how agents really “shop” online\nSurprising findings: agents ignore ads, but positioning still matters\nWhy MCP isn’t enough, and what’s next for protocols\nImplications for retailers, brands, and the future of retail media\nIf you’ve ever wondered how agents will reshape retail and what brands need to do right now to prepare, this episode will give you a front-row seat to the future of agentic commerce.\nHighlights/Timestamps ⏱️\n0:00 – How agents drive omnichannel growth beyond e-commerce\n5:00 – Meet guest Amine Allouah: from École Polytechnique to Columbia PhD\n9:30 – Game theory, algorithm design, and AI: lessons from his PhD\n13:45 – From Meta AI researcher to startup founder: the origin of My Custom AI\n18:10 – What My Custom AI does: feasibility studies, custom model training & workshops\n23:25 – Why retailers sometimes need their own LLMs: accuracy, privacy & cost savings\n29:20 – Beyond LLaMA: multimodal models, recommendation systems, and custom architectures\n34:05 – The ACES Framework: building a sandbox to study agent shopping\n39:16 – Optimizations around pricing, loyalty & group buying\n39:24 – Why...","thumbnail_url":"https://img.transistorcdn.com/bqhboMepVedfOx69KaIjUfykg6rnBbElk4Jko7g1tno/rs:fill:0:0:1/w:400/h:400/q:60/mb:500000/aHR0cHM6Ly9pbWct/dXBsb2FkLXByb2R1/Y3Rpb24udHJhbnNp/c3Rvci5mbS8wMmJh/ZWFmYmNjOTBlMWI0/YzQ5N2VmMTAwYzdh/ZDg1Yi5qcGc.webp","thumbnail_width":300,"thumbnail_height":300}