With the annual celebration of machine learning that is NeurIPS underway, we thought it would be fun to celebrate the amazing accomplishments of our own Hazy alumni! π€©
We're constantly amazed by what they have accomplished -- from landing on the inaugural Time's 100 Most Influential People in AI (2x!) to founding unicorn startups (several!) to winning NSF CAREER Awards (4x!) to earning a spot on Forbes 30 Under 30 (3x!) and too many orals, spotlights, and best papers to count, our former lab members have been busy leaving their mark on the AI landscape.
But beyond their individual accomplishments, what makes us most proud is the work theyβve done to shape the future of AI as professors, entrepreneurs, researchers, and engineers. So please join us in celebrating their amazing accomplishments!
Table of contents:
Albert Gu
Cartesia | CSO
CMU | Assistant Professor
What can we say that Time didnβt already cover in its profile of Albert as part of the inaugural Time Top 100 in AI? Before joining CMU as an assistant professor next year, Albert has been leading the SSM revolution while serving as Chief Scientific Officer of Cartesia.
Notable Papers
- Mamba: Linear-Time Sequence Modeling with Selective State Spaces | COLM 2024
- Efficiently Modeling Long Sequences with Structured State Spaces
- π Oral @ ICLR 2022
- Hippo: Recurrent memory with optimal polynomial projections
- π Spotlight @ NeurIPS 2020
- Kaleidoscope: An efficient, learnable representation for all structured linear maps
- π Spotlight @ ICLR 2020
Awards
- 2024 Timeβs 100 Most Influential People in AI
- 2023 Amazon Research Award
Alex Ratner
Snorkel AI | CEO
University of Washington | Affiliate Assistant Professor
Currently on leave from his lab at UW to continue leading Snorkel AI on its journey from humble academic research project to multi-billion dollar company making data labeling 100x faster for enterprises!
Arjun Desai
Cartesia | Co-Founder
Arjun was recently selected to be on Forbes 30 Under 30 and is helping take SSMs to the moon at Cartesia after graduating from the lab last year! π
Armin Thomas
Liquid AI | Technical Staff
Joined Liquid AI as a technical staff member working on novel ML architectures.
Arun Kumar
RapidFire AI | Cofounder and CTO
University of California, San Diego | Associate Professor (on leave)
Arun has been piling up awards over the past few years for his work on DL systems, and the apple doesnβt fall far from the tree -- his first PhD student, Supun Nakandala, won the 2023 ACM SIGMOD Jim Gray Doctoral Dissertation Award.
Awards
- 2024 VLDB Early Career Research Contribution Award
- 2024 ACM SIGMOD Distinguished Associate Editor
- 2021: NSF CAREER Award
- 2021: IEEE TCDE Rising Star Awards
Beidi Chen
CMU | Assistant Professor
RapidFire AI | Research Scientist
Started the InfiniAI lab at CMU, selected as a Rising Star in EECS by MIT, and just had two papers accepted to NeurIPS 2024 as spotlights! π Beidi is currently looking for new students -- if youβre applying to the CMU PhD program, please reach out to her!
Notable Papers
- Learn To be Efficient: Build Structured Sparsity in Large Language Models | NeurIPS 2024
- π Spotlight @ NeurIPS 2024
- Sequoia: Scalable, Robust, and Hardware-aware Speculative Decoding | NeurIPS 2024
- π Spotlight @ NeurIPS 2024
- MONGOOSE: A Learnable LSH Framework for Efficient Neural Network Training | ICLR 2021
- π Oral @ ICLR 2021
- FlexGen: High-Throughput Generative Inference of Large Language Models with a Single GPU | ICML 2023
- π Oral @ ICML 2023
- Decentralized training of foundation models in heterogeneous environments | NeurIPS 2022
- π Oral @ NeurIPS 2022
Awards
- 2022 ICML Outstanding Paper Runner Up
- 2021 MIT EECS Rising Stars
- 2019 UIUC EECS Rising Stars
Braden Hancock
Meta | Director of AI
After co-founding Snorkel AI, found a new home this year at Meta as the Director of AI, leading the centralized Evaluations org responsible for evaluation strategy, development, tooling, and execution for all GenAI foundation models (e.g. Llama).
Brandon Yang
Cartesia | Co-Founder
After co-founding Cartesia, also earned a spot on Forbes 30 Under 30 this year!
Ce Zhang
University of Chicago | Associate Professor
Together AI | CTO
When not serving as the CTO of one of the fastest growing VC-backed unicorns (Together AI), Ce runs a lab at the University of Chicago focused on distributed learning and efficient inference.
Notable Papers
- Decentralized training of foundation models in heterogeneous environments | NeurIPS 2022
- π Oral @ NeurIPS 2022
Awards
- Co Editor-in-Chief of the new publication DMLR
- 2018 Google Focused Research Award
Christopher De Sa
Cornell | Professor
Together | Advisor
Chris recently received tenure as an associate professor at Cornell while leading the Relax ML Lab π, staying just one step ahead of his amazing students like Ruqi and Cooper whoβve joined the faculty at Purdue and Yale. Chris recently served as Program Chair of MLSys 2024 and won a DARPA grant for decentralized fine-tuning of compressed foundation models.
Notable Papers
- Is My Prediction Arbitrary? The Confounding Effects of Variance in Fair Classification Benchmarks
- π Best Paper Award Honorable Mention for the AI for Social Impact track @ AAAI 2024
- QuIP: 2-Bit Quantization of Large Language Models With Guarantees
- π Spotlight @ NeurIPS 2023
- Low-Precision Stochastic Gradient Langevin Dynamics
- π Spotlight @ ICML 2022
- A General Analysis of Example-Selection for Stochastic Gradient Descent
- π Spotlight @ ICLR 2022
- Optimal Complexity in Decentralized Training
- π Oral @ ICML 2021
- Asymptotically Optimal Exact Minibatch Metropolis-Hastings
- π Spotlight @ NeurIPS 2020
- Random Reshuffling is Not Always Better
- π Spotlight @ NeurIPS 2020
- Regulating Accuracy-Efficiency Trade-Offs in Distributed Machine Learning Systems
- π Oral @ LML 2020: ICML Workshop on Law and Machine Learning
- Numerically Accurate Hyperbolic Embeddings Using Tiling-Based Models
- π Spotlight @ NeurIPS 2019
- Distributed Learning with Sublinear Communication Long oral
- π Oral @ ICML 2019
- Poisson-Minibatching for Gibbs Sampling with Convergence Rate Guarantees
- π Spotlight @ NeurIPS 2019
Awards
- 2024 keynote at the International Conference on AI-ML Systems
- 2024 DARPA grant - decentralized fine-tuning of compressed foundation models
- 2021 Google Research Scholar Award
- 2021 NSF CAREER Award
Dan Fu
UCSD | Assistant Professor
Together AI | Research Scientist
Pending his ability to fulfill basic PhD requirements, Dan will soon join UCSD as the newest member of the computer science faculty π!
Also recently adopted a new kitten!
Notable Papers
- FlashFFTConv: Efficient Convolutions for Long Sequences with Tensor Cores
- π Oral @ ββENLSP Workshop at NeurIPS 2023
- Monarch Mixer: A Simple Sub-Quadratic GEMM-Based Architecture
- π Oral @ NeurIPS 2023
- Hungry hungry hippos: Towards language modeling with state space models
- π Spotlight @ ICLR 2023
- FlashAttention: Fast and memory-efficient exact attention with IO-awareness
- π Best Paper @ Hardware Aware Efficient Training Workshop at ICML 2022
- RedPajama: an Open Dataset for Training Large Language Models
- π Spotlight @ NeurIPS Datasets 2024
- Hyena Hierarchy: Towards Larger Convolutional Language Models
- π Oral @ ICML 2023
- Shoring Up the Foundations: Fusing Model Embeddings and Weak Supervision
- π Oral @ UAI 2022
Awards
- Stanford Open Source Software Prize 2024
- NDSEG Award for Exemplary Impact and Relevance
Feng Niu
Evidently | CEO
Founded Evidently to empower clinicians to make better decisions, leading to happier clinicians, healthier patients, and improved financial ROI for hospitals
Fred Sala
UW Madison | Assistant Professor
Snorkel AI | Chief Scientist
When not serving as Chief Scientist of Snorkel AI, Fred has kept himself busy winning the 2024 DARPA Young Faculty Award for work on data-efficient decentralized foundation models and leading his lab at UW Madison focused on developing data-driven systems for machine learning.
Notable Papers
- The AlCHEmist: Automated Labeling 500x CHEaper than LLM Data Annotators | NeurIPS 2024
- π Spotlight @ NeurIPS 2024
- Promises and Pitfalls of Threshold-based Auto-labeling | NeurIPS 2023
- π Spotlight @ NeurIPS 2023
- Skill-it! A Data-Driven Skills Framework for Understanding and Training Language Models | NeurIPS 2023
- π Spotlight @ NeurIPS 2023
- Domain Generalization via Nuclear Norm Regularization | CPAL 2023
- π Spotlight @ CPAL 2023
- Foundation Models Can Robustify Themselves, For Free | Neurips 2023
- π Best Paper Award Honorable Mention @ NeurIPS R0-FoMo Workshop 2023
- Shoring Up the Foundations: Fusing Model Embeddings and Weak Supervision
- π Oral @ UAI 2022
Awards
- 2024 DARPA Young Faculty Award
Ines Chami
Numbers Station | Cofounder
Graduated from the lab π and cofounded Numbers Station to apply LLMs to structured data; before that, won the 2021 Stanford Gene Golub Doctoral Dissertation Award!
Jaeho Shin
Evidently | Chief Engineer
Cofounded Evidently to build tools that doctors and nurses love to use to improve the care they deliver to patients
Jared Dunnmon
Chief Scientist | Stealth Startup
Former Technical Director for AI | Defense Innovation Unit
Served as the first Technical Director for Artificial Intelligence at the Defense Innovation Unit (DIU), during which he ran the DIU xView3 program (a public challenge resulting in ML models currently used to fight illegal fishing around the world) and built out the Emerging Capabilities Policy Office at the Pentagon. Recently cofounded a company focused on rebuilding the U.S. maritime industry.
Jason Fries
Research Scientist | Stanford University
Snorkel AI | Advisor
Working at the intersection of multimodal foundation models and healthcare, Jason has developed methods and datasets for unlocking the full potential of generative AI in real-world medical settings.
Notable Papers
- MedAlign: A Clinician-Generated Dataset for Instruction Following with Electronic Medical Records
- π Best Thematic Paper Award @ AHLI Machine Learning for Health Symposium 2023
- MOTOR: A Time-To-Event Foundation Model For Structured Medical Records
- π Spotlight @ ICLR 2024
- EHRSHOT: An EHR Benchmark for Few-Shot Evaluation of Foundation Models
- π Spotlight @ NeurIPS Datasets 2023
Awards
- 2021 HAI-AIMI Partnership Grant for developing multimodal patient embeddings
Karan Goel
Cartesia | CEO
Taking SSMs to the moon as CEO of Cartesia after graduating from the lab last year! π Currently looking for smart ML researchers (preferably named Brandon)
Khaled Saab
Google DeepMind | Research Scientist
Graduated from the lab π before joining Google DeepMind to push the state-of-the-art of using foundation models to improve healthcare through leading efforts like Med-Gemini and AMIE.
Notable Papers
- Med-Gemini: Capabilities of Gemini Models in Medicine
- AMIE: Towards Conversational Diagnostic AI
- Hungry hungry hippos: Towards language modeling with state space models
- π Spotlight @ ICLR 2023
- Spatiotemporal Modeling of Multivariate Signals With Graph Neural Networks and Structured State Space Models
- π Best Paper @ CHIL 2023
- Domino: Discovering Systematic Errors with Cross-Modal Embeddings
- π Oral @ ICLR 2022
- Doubly Weak Supervision of Deep Learning Models for Head CT
- π Oral @ MICCAI 2019
- Improving Sample Complexity with Observational Supervision
- π Spotlight @ ICLR 2019 Limited Labeled Data Workshop
Kun-Hsing Yu
Harvard Medical School | Assistant Professor
Published many amazing papers with his lab on AI and digital pathology, developed numerous tools for cancer diagnosis and treatment, and co-founded a biotech company for cancer pathology diagnosis.
Notable Papers
- A Pathology Foundation Model for Cancer Diagnosis and Prognostic Prediction | Nature 2024
- π Nature
- Medical Artificial Intelligence and Human Values | NEJM 2024
- π NEJM
Awards
- NIH R01 and R35 grants
- Department of Defense (DoD) Career Development Award
- American Cancer Society Research Award
- Google Research Scholar Award
Kush Bhatia
Google Deepmind | Research Scientist
Finished his postdoc at Hazy Research π and joined Google Deepmind
Notable Papers
- TART: A plug-and-play Transformer module for task-agnostic reasoning
- π Spotlight @ R0-FoMo at NeurIPS 2023
- Skill-it! A data-driven skills framework for understanding and training language models
- π Spotlight @ NeurIPS 2023
- On the Sensitivity of Reward Inference to Misspecified Human Models
- π Spotlight @ ICLR 2023
Awards
- JP Morgan PhD Fellowship 2020
Nancy Xu
Moonhub | CEO
Xu Ventures | Founder, General Partner
Another Time 100 AI honoree! Built the world's first AI Recruiter as the founder of Moonhub and is a WEF Technology Pioneer
Awards
- 2024 Timeβs 100 Most Influential People in AI
- WEF Technology Pioneer
Nimit Sohoni
Cartesia | Researcher
Joined Cartesia as an early member of the research staff after working at Citadel for 2 years!
Sarah Hooper
National Institutes of Health | Research Scientist
Graduated from the lab! π Decided being an academic didnβt involve enough bureaucracy and is now a research scientist at the NIHβs Office of AI Research.
Notable Papers
- A case for reframing automated medical image classification as segmentation | NeurIPS 2023
- Fine-Grained Vision-Language Representation Learning from Real-World Data | ICCV 2023
- Reducing Reliance on Spurious Features in Medical Image Classification with Spatial Specificity | ML4H 2022
Awards
- 2023 Human-Centered AI and GCP Research Proposal
- Hertz Foundation Fellowship
Sharon Li
UW-Madison | Assistant Professor
Recently named β2023 Innovator of the Yearβ by MIT Technology Review after running out of other awards to win. Her lab at UW Madison focuses on algorithmic and theoretical foundations of reliable machine learning, addressing challenges in both model development and deployment in the open world!
Notable Papers
- HaloScope: Harnessing Unlabeled LLM Generations for Hallucination Detection
- π Spotlight @ NeurIPS 2024
- A Graph-Theoretic Framework for Understanding Open-World Semi-Supervised Learning
- π Spotlight @ NeurIPS 2023
- Distributionally Robust Optimization with Probabilistic Group
- π Oral @ AAAI 2023
- Out-of-distribution Detection via Frequency-regularized Generative Models
- π Spotlight @ WACV 2023
- Is Out-of-Distribution Detection Learnable?
- π Outstanding Paper Award @ NeurIPS 2022
- POEM: Out-of-Distribution Detection with Posterior Sampling
- π Oral @ ICML 2022
- Unknown-Aware Object Detection: Learning What You Donβt Know from Videos in the Wild
- π Oral @ CVPR 2022
- PiCO: Contrastive Label Disambiguation for Partial Label Learning
- π Outstanding Paper Award Honorable Mention @ ICLR 2022
- Provable Guarantees for Understanding Out-of-distribution Detection
- π Oral @ AAAI 2022
- On the Impact of Spurious Correlation for Out-of-distribution Detection
- π Oral @ AAAI 2022
Awards
- 2023 Innovator of the Year by MIT Technology Review
- 2023 NSF CAREER Award
- 2022 AFOSR Young Investigator Program (YIP) Award
- 2022 Amazon Research Award
- 2021 Google-Initiated Focused Research Award
- 2021 Facebook Research Award
- 2021 JP Morgan Early-career Faculty Award
- 2020 Forbes 30 Under 30
- 2019 30 Under 30 Rising Stars in AI
Silas Alberti
Cognition AI | Founding Engineer
Automating himself out of a job while building Devin as a founding engineer at Cognition AI
Stephen Bach
Brown | Assistant Professor
Snorkel AI | Advisor
In addition to leading the BATS research group at Brown, co-led one of the teams that proposed instruction tuning, released the T0 family of models, and achieved top performance on zero-shot image classification in DARPA's Learning with Less Labeling program.
Notable Papers
- T0: Multitask prompted training enables zero-shot task generalization
- π Spotlight @ ICLR 2022
- Low-Resource Languages Jailbreak GPT-4
- π Best Paper Award @ NeurIPS Workshop on Socially Responsible Language Modeling Research 2023
- Snorkel: Rapid training data creation with weak supervision
- π Best of VLDB 2018
Awards
- Co-PI on DARPA grant βTasks Algorithmically Given Labels Established via Transferred Symbolsβ
- NSF grant to develop an LLM-backed research assistant for comprehensive climate assessments
Theo Rekatsinas
Apple | Research Scientist
Currently at Apple enabling smaller generative models to have the capabilities of frontier models via data selection methods after his data cleaning company, Inductiv, was acquired. In a previous life, was a Professor at UW Madison and ETH working on a $5M DARPA grant for scientific discovery and semantic analytics
Tri Dao
Princeton | Assistant Professor
Together AI | Chief Scientist
Starting a new lab at Princeton next year π after doing a stint as Chief Scientist of Together AI! Made transformers significantly more scalable with FlashAttention 1/2/3, then decided to burn it all down with Mamba.
Notable Papers
- Mamba: Linear-Time Sequence Modeling with Selective State Spaces | COLM 2024
- FlashAttention: Fast and memory-efficient exact attention with IO-awareness
- π Best Paper @ Hardware Aware Efficient Training Workshop at ICML 2022
- Hyena hierarchy: Towards larger convolutional language models
- π Oral @ ICML 2023
- Monarch: Expressive structured matrices for efficient and accurate training
- π Outstanding Paper runner-up @ ICML 2022
- Decentralized training of foundation models in heterogeneous environments
- π Oral @ NeurIPS 2022
- Hungry hungry hippos: Towards language modeling with state space models
- π Spotlight @ ICLR 2023
- Pixelated butterfly: Simple and efficient sparse training for neural network models
- π Spotlight @ ICLR 2022
- Kaleidoscope: An efficient, learnable representation for all structured linear maps
- π Spotlight @ ICLR 2020
- Hippo: Recurrent memory with optimal polynomial projections
- π Spotlight @ NeurIPS 2020
Awards
- Stanford Open Source Software Prize 2024
Vincent Chen
Snorkel AI | Founding Engineer, VP
Still working on Snorkel AI!
Virginia Smith
Carnegie Mellon University | Associate Professor
When not busy running her lab at CMU focused on federated learning and privacy-preserving ML, won an NSF CAREER Award and Sloan Research Fellowship and came in 1st place in the 2023 UK-US Privacy-Enhancing Technologies pandemic forecasting challenge
Notable Papers
- Fair Federated Learning via Bounded Group Loss
- π Best Paper Award @ ICLR 2022 Socially Responsible ML Workshop
- Validating Large Language Models with ReLM
- π Outstanding Paper Award @ MLSys 2023
- Ditto: Fair and Robust Federated Learning Through Personalization
- π Best Paper Award @ ICLR 2021 Secure ML Workshop
Awards
- NSF CAREER Award
- Sloan Research Fellowship
- 2023 Samsung AI Researcher of the Year
- 2022 Intel Rising Star
- MIT Technology Review's 35 Innovators Under 35
- 1st place in the 2023 UK-US Privacy-Enhancing Technologies pandemic forecasting challenge
If you're an alum and want to be added to this list or modify your information, please shoot Michael (mwornow@stanford.edu) an email or Slack message!