Scholar
Dan Biderman
Google Scholar ID: 6WFbVJUAAAAJ
Stanford University
Citations & Impact
Citations
597
H-index
8
i10-index
8
Publications
20
Co-authors
7
list available
Publications
Resume (English only)
Academic Achievements
- “LoRA Learns Less and Forgets Less” published in TMLR 2024 (Featured Certification)
- “Minions: Cost-efficient Collaboration Between On-device and Cloud Language Models” accepted at ICML 2025
- Led development of Lightning Pose package; paper published in Nature Methods 2024
- “Partitioning variability in animal behavioral videos using semi-supervised variational autoencoders” published in PLoS Computational Biology 2021
- Recipient of Columbia’s Titus M Cowan Dissertation Prize in Biomedical Research
- Student speaker at Columbia’s 2025 PhD hooding ceremony
Research Experience
- Postdoc at Linderman Lab (Stanford Statistics) and Hazy Research (Ré) Lab (Stanford CS)
- Developed deep learning models for animal pose tracking in videos during PhD at Columbia’s Center for Theoretical Neuroscience (Lightning Pose project)
- Collaborated with Databricks Mosaic AI on learning-forgetting tradeoffs in parameter-efficient finetuning
- Proposed new collaboration patterns between on-device and cloud LLMs (Minions project)
- Co-organizes the workshop on Efficient Systems for Foundation Models (most recently at ICML 2025)
Background
- Postdoctoral Scholar at Stanford University, jointly affiliated with Statistics and Computer Science
- Co-advised by Christopher Ré and Scott Linderman
- Builds resource-efficient AI systems and applies them to neuroscience
- Current focus: language models that dynamically learn from experience
- Research interests: Efficient LLMs and multi-agent systems; hardware-aware numerical linear algebra and ML; modeling and analysis of biological data
