Nathan Lambert

Nathan Lambert is an artificial intelligence researcher, author and leading voice on open language models, reinforcement learning from human feedback and LLM post-training. A former Senior Research Scientist at the Allen Institute for AI, Nathan is particularly recognised for spearheading the alignment and instruction-tuning interface as the post-training lead for AI2’s OLMo initiative. As an AI speaker, he brings audiences direct insight into how advanced language models are trained, aligned and developed, making him ideal for technology conferences, leadership summits and events exploring the future of generative AI.

Nathan’s technical foundations span engineering, robotics and artificial intelligence. At Cornell University’s SonicMEMS Laboratory, he investigated micro-electro-mechanical systems for chip-scale gas sensing and supported Tesla Motors’ Battery Engineering team during the Model X launch. He later joined Facebook AI Research as a student researcher, focusing on robotics and machine learning control algorithms, before working as a Research Scientist Intern at Google DeepMind. He earned his Ph.D. in Electrical Engineering and Computer Sciences from UC Berkeley, where he also co-instructed CS 188: Artificial Intelligence to more than 800 students. This combination of research and teaching enables Nathan to communicate complex AI concepts with technical authority and clarity.

Nathan subsequently built and led Hugging Face’s dedicated Reinforcement Learning from Human Feedback research team and founded Interconnects AI, growing its technical frontier AI newsletter to more than 50,000 subscribers. He later joined AI2 to research RLHF and LLM fine-tuning before playing a central role in OLMo. He has since founded a stealth AI laboratory focused on open-source language models and published ‘Reinforcement Learning from Human Feedback: Reinforcement Learning from Human Feedback, Alignment, and Post-training LLMs’.

Nathan’s influence through Interconnects AI and his work across internationally recognised AI organisations give event audiences a perspective grounded in both research and real-world model development. His experience spans the organisations shaping modern AI, from Google DeepMind and Facebook AI Research to Hugging Face and AI2. Organisations hire Nathan as an AI speaker for informed, technically rigorous perspectives on RLHF, post-training, AI alignment and open-source models. He can explain not only where frontier AI is heading, but how these systems are actually built and improved, giving technical audiences, executives and innovators practical insight into the decisions shaping the next generation of artificial intelligence.