Jonathan Frankle

Chief Scientist at MosaicML (acquired by Databricks)

Leading Mosaic Research lab, advancing AI training efficiency and neural network learning.

Training an LLMs with the hundreds-of-mini-cupcakes approach

MosaicML was just acquired for $1.3 Billion.

Jonathan Frankle serves as the Chief Neural Network Scientist at Databricks, leading the Mosaic Research lab comprised of over 30 research scientists. With a focus on enhancing the efficiency of training modern generative AI models like LLMs and diffusion models, his team conducts empirical studies on neural network learning.

Jonathan’s journey to Databricks followed the company’s $1.3B acquisition of MosaicML, where he played a pivotal role as a founding team member. He holds a PhD in computer science from MIT, with his research contributions accessible on Google Scholar. Based in New York City, Jonathan frequently travels between San Francisco and Washington, DC, advancing the frontier of AI research.