Jonathan Frankle

Also known as jefrankle

Jonathan Frankle is a deep learning researcher who, with Michael Carbin, introduced the lottery ticket hypothesis: dense networks contain subnetworks, called winning tickets, that train effectively. He studied computer science at Princeton, served as the inaugural staff technologist at Georgetown Law's Center on Privacy & Technology in 2015–2016, and completed a PhD at MIT in 2023. The lottery ticket paper won an ICLR 2019 Best Paper Award, and his thesis received the 2023 AAAI/ACM SIGAI Doctoral Dissertation Award. MosaicML, founded in 2021 with him as chief scientist, was acquired by Databricks in 2023, and in March 2024 he co-authored the announcement of the open-source DBRX model. He is Chief AI Scientist at Databricks, where he leads research on reinforcement learning, model training and agent evaluation.

Now at: Databricks
Badges: Bridge

Timeline

2015–2016

Staff technologist at Georgetown Law

In 2015–2016, he is the inaugural staff technologist at the Center on Privacy & Technology at Georgetown Law. There he contributes to The Perpetual Lineup, a report on police use of face recognition.
2016–2023

A PhD at MIT

After earning bachelor's and master's degrees in computer science at Princeton, he begins a PhD at MIT in 2016, where he studies deep learning with Professor Michael Carbin. He completes the degree in 2023.
2018

The lottery ticket hypothesis

In March 2018, he and Michael Carbin post "The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networks" on arXiv. They find that a standard pruning technique uncovers subnetworks of dense networks whose initializations make them capable of training effectively, and they call these subnetworks winning tickets.
2019

ICLR 2019 best paper award

The lottery ticket paper wins an ICLR 2019 Best Paper Award, shared with "Ordered Neurons: Integrating Tree Structures into Recurrent Neural Networks" by Yikang Shen, Shawn Tan, Alessandro Sordoni and Aaron Courville.
2021–2023

Chief scientist at MosaicML

MosaicML is founded in 2021, with him as chief scientist. Databricks acquires the company in 2023, in a deal valued at roughly $1.3 billion.
2023

A SIGAI dissertation award

He receives the 2023 AAAI/ACM SIGAI Doctoral Dissertation Award for his MIT thesis, "The Lottery Ticket Hypothesis: On Sparse, Trainable Neural Networks."
March 2024

The DBRX language model

In March 2024, Databricks releases DBRX, an open-source general-purpose language model built by its AI research team, and he co-authors the announcement. He says DBRX matters because Databricks tools built it, so any customer can reach similar performance with their own models.
As of October 2026

Chief AI Scientist at Databricks

He is Chief AI Scientist at Databricks, where he leads the research team and its projects on reinforcement learning, model training and agent evaluation.