# Tim Dettmers

> Tim Dettmers is an Assistant Professor at Carnegie Mellon University and Research Scientist at the Allen Institute for AI (AI2), best known for making large language models accessible on consumer hardware. He created the bitsandbytes library (2.2M monthly installs), co-authored QLoRA - a technique enabling fine-tuning of 65B-parameter models on a single GPU - and pioneered LLM.int8() quantization. With over 18,000 citations across his work, Dettmers has become one of the most influential voices in efficient deep learning, consistently arguing that computational democratization - not AGI hype - is where the real progress lives.

- **Role:** Assistant Professor at Carnegie Mellon University; Research Scientist at Allen Institute for AI (AI2)
- **Organizations:** Carnegie Mellon University, Allen Institute for AI (AI2), bitsandbytes-foundation
- **From:** Germany
- **Nationality:** German
- **Education:** PhD, Computer Science, University of Washington
- **Known for:** Created bitsandbytes library with 2.2 million monthly installations, Co-authored QLoRA - enabling fine-tuning of 65B parameter models on a single consumer GPU, Pioneered LLM.int8() - first practical 8-bit quantization for multi-billion parameter transformers

## Career timeline

- **Pre-2018** — Spent 3 years in the automation industry in Germany developing SCADA systems
- **2016** — Received Google Scholarship
- **2018** — Began PhD at University of Washington under Luke Zettlemoyer; received Jeff Dean - Heidi Hopper Endowed Regental Fellowship
- **2021** — Won NeurIPS Best Reviewer Award
- **2022** — Published LLM.int8() - 8-bit quantization enabling multi-billion parameter model inference on consumer GPUs
- **2023** — Co-authored QLoRA paper enabling 65B model fine-tuning on a single 48GB GPU; received Google Open Source Award and PyTorch Foundation Award
- **2024** — Completed PhD; joined Allen Institute for AI as Research Scientist; named AI2050 Early Career Fellow (Schmidt Sciences); published SpQR at ICLR 2024
- **2025** — Started as Assistant Professor at Carnegie Mellon University (School of Computer Science); received Google ML and Systems Junior Faculty Award ($100K)
- **2026** — Published SERA research on efficient coding agents; continued research on agent systems and MoEs at CMU and AI2

## Achievements

- Created bitsandbytes library with 2.2 million monthly installations
- Co-authored QLoRA - enabling fine-tuning of 65B parameter models on a single consumer GPU
- Pioneered LLM.int8() - first practical 8-bit quantization for multi-billion parameter transformers
- Published SpQR at ICLR 2024 - near-lossless LLM weight compression
- Over 18,272 total citations across research portfolio
- 2025 Google ML and Systems Junior Faculty Award ($100,000 unrestricted)
- 2024 AI2050 Early Career Fellowship - Schmidt Sciences
- 2023 Google Open Source Award
- 2023 PyTorch Foundation Award
- 2021 NeurIPS Best Reviewer Award
- 2018/2019 Jeff Dean - Heidi Hopper Endowed Regental Fellowship
- 2016/2017 Google Scholarship
- Block Award and Madrona Prize at University of Washington
- SERA coding agent achieves SOTA among fully open-source models at 26x cheaper than RL, 57x cheaper than previous synthetic data methods

## Latest updates

- **2026-01** — Published SERA research - efficient coding agents achieving SOTA results among open-source models at dramatically lower cost than RL approaches
- **2025-12** — Published 'Why AGI Will Not Happen' blog post, covered by The Register and generating significant discussion in AI community
- **2025-09** — Received Google ML and Systems Junior Faculty Award ($100,000 unrestricted research funding)
- **2025-08** — Joined CMU Catalyst for agent systems research focusing on bringing best models to consumer GPUs
- **2025-01** — Started as Assistant Professor at Carnegie Mellon University School of Computer Science
- **2024-11** — Shifted focus at AI2 toward building general agent frameworks
- **2024-08** — Made strategic shift to focus on coding agents and agent systems research

## Links

- Website: https://timdettmers.com/
- LinkedIn: https://us.linkedin.com/in/timdettmers
- Twitter/X: https://x.com/Tim_Dettmers
- GitHub: https://github.com/timdettmers

---

Profile page: https://yespress.io/tim-dettmers
Published by YesPress — https://yespress.io
Last updated: 2026-04-20