Hi! I'm Jacob, and I'm a Ph.D. student at the University of Washington, where I'm advised by Noah Smith and supported by an NSF Computer Science Graduate Fellowship. I'm also a student researcher on the Open Ecosystem team at Ai2, where I mainly work on language models, and an AI Fellow at Madrona Venture Group.
Before my Ph.D., I worked on code generation at Google [x] and platform health at Twitter, and in a previous life I was a software engineer at Tableau and Google. See my CV for more details.
What I work on
Language models
I work broadly on the full model development pipeline, from model architecture to pretraining and post-training. My work generally aims to improve both overall model capabilities and the downstream efficiency and flexibility of state of the art models.
Technology & Society
I also study the impact of AI and other emerging technologies on society writ large. I've studied the environmental cost of building and deploying language models, from hardware manufacturing through deployment and for reasoning models, and I've written about legal questions raised by generative AI, such as liability, legal risk, and the First Amendment.
I now spend some of my time thinking about the environmental impact of data center buildouts, energy policy in the U.S. and Washington State, and the impact of AI deployment on public education and labor.
Selected awards
- May 2025Mercor Graduate Fellowship Finalist
- Aug 2024ACL Theme Paper Award
- Aug 2024ACL Best Resource Paper Award
- Apr 2024GeekWire Innovation of the Year
- Aug 2023NSF Computer Science Graduate Fellowship
Selected publications
See all →* indicates equal contribution
- Olmo 3
- Tülu 3: Pushing Frontiers in Open Language Model Post-Training
- Train Separately, Merge Together: Modular Post-Training with Mixture-of-Experts
-
Holistically Evaluating the Environmental Impact of Creating Language Models
ICLR 2025 paper 🔦 Spotlight Paper 🔦
-
The Hidden Cost of Thinking: Energy Use and Environmental Impact of LMs Beyond Pretraining
COLM 2026 paper
-
Intentionally Unintentional Speech: Why Generative AI Models Are Not Protected by the First Amendment
First Amendment Law Review (University of North Carolina), 2024 paper
Talks & press
- Apr 2026 Talk Ecological Impacts of AI Seattle University School of Law
- Dec 2025 Press Why is Everyone So Wrong About AI Water Use?? Hank Green (YouTube) (over 5.4M views)
- Nov 2025 Talk RewardBench: Evaluating Reward Models for Language Modeling Georgetown Law
- Jul 2025 Press A new kind of AI model gives data owners more control Wired
- Jul 2024 Briefing Technical and Policy Considerations for Generative AI Washington State House Consumer Protection & Business Committee
- Nov 2023 Talk Responsible, Real World AI Development and Deployment State of Washington AI Community of Practice
- Aug 2023 Event Future of AI Forum, organized with U.S. Senator Maria Cantwell's office covered by GeekWire