Dan Biderman
Dan Biderman is a computational neuroscientist, machine learning researcher and co-founder of Engram. He studied visual perception at Tel Aviv University, developed resource-efficient ways of analyzing behavior and training models at Columbia, and pursued postdoctoral research at StanfordStanford UniversityStanford University is a research university in California founded by Leland and Jane Stanford in 1885 and opened in 1891.Open the full entry. He helped develop software for tracking movement in video and went on to study how language models could retain what they learned from work.
Contents6 sections
Key facts
2013–2018Family and study in Tel Aviv
In his dissertation Biderman described a family home in which his parents, Eytan and Sarit, valued both learning and lasting relationships. He named Michael and Gaya as his siblings. His early intellectual interests included consciousness and how humans and other animals perceive the world. From 2013 to 2017 he studied mathematics, neurobiology, cognitive science and the history and philosophy of science in Tel Aviv University’s Adi Lautman Interdisciplinary Program, which led directly to a master’s degree.[1][2][3]
He received a full-tuition merit scholarship for four years and an outstanding achievements award in 2015. He received travel grants for an advanced course on consciousness in Italy in 2016 and for the 2017 Vision Sciences Society meeting in Tampa, FloridaFloridaFlorida is a U.S. state occupying mainly the Florida peninsula, with its capital at Tallahassee; its most populous city is Jacksonville. It entered the Union on March 3, 1845, as the 27th state, and had a population of 21,538,187 in the 2020 census (April 1, 2020).Open the full entry. The social sciences department recognized him as an outstanding graduate student in 2017. He earned an MA in cognitive science in 2018 under Liad Mudrik. His thesis used psychophysics and statistical modeling to examine how context influences the perception of objects that can be interpreted in more than one way.[2]
2018–2024Neuroscience at Columbia
He moved to the United StatesUnited StatesThe United States is a federal country in North America made up of 50 states, the District of Columbia as the seat of government, and territories such as Puerto Rico and Guam; its capital is Washington, D.C. The Continental Congress adopted the Declaration of Independence on July 4, 1776. The 2020 census (April 1, 2020) counted a resident population of 331,449,281, and the 435 seats of the House of Representatives are apportioned among the states on the basis of the census.Open the full entry for graduate study at Columbia, earning an MSc in neurobiology and behavior in 2020 and a PhD in the same field in 2024. His adviser John Cunningham taught him probabilistic modeling. Liam Paninski became an informal co-adviser during his first rotation and supported him during the move to the United States. His dissertation examined resource-efficient machine learning, from animal movement in video to natural language. Working with Matthew Whiteway and others, he helped develop Lightning Pose to track body positions in recorded behavior. The work appeared in Nature Methods in 2024.[2][1][4][5]
He described starting a family while building his research career. His wife Natalie gave detailed feedback on his writing, figures and presentations. His in-laws Simha and Raisa supported the family, and Cunningham passed along equipment for the children. Biderman dedicated the dissertation to Natalie and their children Ella and Ram.[1]
2023–2025From an industry internship to a Stanford postdoc
From May to August 2023 he interned with MosaicML’s natural language processing data team. From November 2023 through June 2024 he worked part time with the MosaicML and Databricks team on tuning models for code and mathematics with fewer trainable parameters. His work on LoRA examined the trade-off between learning new material and forgetting existing capabilities. He was also a Lightning AI community ambassador.[2][5][6]
At Columbia he assisted in teaching theoretical neuroscience, gave guest lectures on modeling behavior in a neural data analysis course and mentored two undergraduates. He declined a 2024 research fellowship at the Simons Foundation’s Flatiron Center for Computational Mathematics. In September of that year he began postdoctoral work in statistics and computer science at Stanford under Christopher Ré and Scott Linderman. The Linderman lab records his appointment as 2024–25. In the Minions project, presented at ICML 2025, he and his collaborators studied how to share work between language models on a device and in the cloud. He also co-organized a workshop on efficient systems for foundation models. Columbia awarded him the Titus M. Coan dissertation prize for excellence in biomedical research and selected him to speak for graduating students at its 2025 doctoral hooding ceremony.[2][7][5][8]
2026Engram and memory acquired at work
Biderman co-founded Engram with Jessy LinJessy LinJessy Lin is a computer scientist and a co-founder of Engram. She attended high school on Long Island, studied electrical engineering and computer science and philosophy at MIT, and earned a PhD in computer science at the University of California, Berkeley. Her work concerns how language models can learn from feedback, retain experience and make decisions with people.Open the full entry and other researchers and became its CEO. The company announced its first product for shared workspaces and described partnerships with Notion and the legal technology company Harvey. MicrosoftMicrosoftMicrosoft is a technology company founded by Bill Gates and Paul Allen in 1975 and headquartered in Redmond, Washington, with businesses spanning Windows, Microsoft 365, Azure, enterprise software, artificial intelligence, gaming and devices.Open the full entry also partnered with Engram on Microsoft 365. Engram reported raising $98 million. In a June 24 Sequoia interview with Lin, Biderman offered the example of sticky notes. People consult written records, yet return to work with intuitions formed the day before. He argued that designers should decide what a system needs to learn, what it can look up externally, and how it can work with the tools particular to each team.[9][3][10]
Related Organizations, Websites and Public Accounts
MosaicML: Research intern and subsequent part-time researcher with the Databricks team. Official website: https://www.databricks.com/ (opens in a new window).[2][6]
Stanford University: Former postdoctoral scholar in statistics and computer science. Official website: https://www.stanford.edu/ (opens in a new window).[7][2]
Engram: Co-founder and CEO. Official website: https://engram.com/ (opens in a new window).[9][10][3]
Personal website: https://dan-biderman.netlify.app/ (opens in a new window). GitHub: https://github.com/danbider (opens in a new window). X: https://x.com/dan_biderman (opens in a new window). LinkedInLinkedInLinkedIn is a social networking platform for professionals. It was founded in 2002 by Reid Hoffman and others, launched in 2003, and listed on the New York Stock Exchange in 2011; after Microsoft acquired it for approximately $26.2 billion in 2016, it became a wholly owned Microsoft subsidiary. Its revenue comes from membership subscriptions, advertising, and recruitment solutions.Open the full entry: https://www.linkedin.com/in/dan-biderman-066039202 (opens in a new window).[5][11][10]
Sources11 sources
- Resource-Efficient Machine Learning Systems: From Natural Behavior to Natural Language (opens in a new window)academiccommons.columbia.edu↑
- Dan Biderman, Curriculum Vitae (opens in a new window)dan-biderman.netlify.app↑
- Memory and Continual Learning: Engram's Dan Biderman and Jessy Lin (opens in a new window)youtube.com↑
- Alumni: Doctoral Program in Neurobiology and Behavior (opens in a new window)neurosciencephd.columbia.edu↑
- Dan Biderman (opens in a new window)dan-biderman.netlify.app↑
- Dan Biderman (opens in a new window)developer.nvidia.com↑
- Linderman Lab at Stanford University (opens in a new window)lindermanlab.github.io↑
- Student Handbook 2025–26 (opens in a new window)dbmi.columbia.edu↑
- Engram: AI That Learns From You (opens in a new window)engram.com↑
- Dan Biderman (opens in a new window)linkedin.com↑
- Dan Biderman (opens in a new window)github.com↑