Danqi Chen

Career

Chen received a PhD in computer science from Stanford University in 2018, advised by Christopher Manning in the Stanford NLP Group; the thesis later received Stanford's Arthur Samuel Best Doctoral Thesis Award. At Princeton, Chen is an associate professor of computer science, co-leads the Princeton NLP Group, and is an associate director of Princeton Language and Intelligence.¹⁵

Outside Princeton, Chen has worked as a research scientist at Facebook AI Research and Google DeepMind, and spent the 2025–26 academic year as a member of technical staff at Thinking Machines Lab.¹

Research

Chen's earlier work focused on reading comprehension and open-domain question answering, including first-authored work on answering open-domain questions by reading Wikipedia (2017) and co-authorship of Dense Passage Retrieval (2020). Chen also co-authored the RoBERTa study of BERT pretraining in 2019 and SimCSE, a contrastive method for sentence embeddings, in 2021.⁶⁷²³

Chen's more recent papers span language model training and alignment, long-context modeling, pretraining data curation, interpretability, and formal theorem proving. Examples include SimPO, a reference-free preference optimization method at NeurIPS 2024, and Goedel-Prover-V2, a series of open-source theorem-proving models at ICLR 2026.⁸⁹

Work and contributions

• SimCSE Open-source code

  Co-author with Tianyu Gao and Xingcheng Yao of the EMNLP 2021 SimCSE paper on contrastive learning of sentence embeddings; the princeton-nlp/SimCSE repository, released under the MIT license, hosts the code and lists Chen in its citation.³⁴

• RoBERTa: A Robustly Optimized BERT Pretraining Approach Published research

  Listed among the authors of the 2019 RoBERTa paper, a replication study of BERT pretraining.²

• Reading Wikipedia to Answer Open-Domain Questions Published research

  First author, with Adam Fisch, Jason Weston and Antoine Bordes, of the 2017 paper on answering open-domain questions using Wikipedia as the knowledge source.⁶

• SimPO: Simple Preference Optimization with a Reference-Free Reward Published research

  Last author, with Yu Meng and Mengzhou Xia, of this NeurIPS 2024 paper on a preference optimization method that does not need a reference model.⁹⁸

Timeline

1. Mar 31, 2017Posted Reading Wikipedia to Answer Open-Domain Questions as first author, with Adam Fisch, Jason Weston and Antoine Bordes.⁶
2. 2018Completed a PhD in computer science at Stanford University, advised by Christopher Manning.¹
3. Jul 26, 2019Co-authored RoBERTa: A Robustly Optimized BERT Pretraining Approach.²
4. Apr 10, 2020Co-authored Dense Passage Retrieval for Open-Domain Question Answering.⁷
5. Apr 18, 2021Co-authored SimCSE: Simple Contrastive Learning of Sentence Embeddings with Tianyu Gao and Xingcheng Yao.³
6. May 23, 2024Co-authored SimPO: Simple Preference Optimization with a Reference-Free Reward, later at NeurIPS 2024.⁹⁸
7. 2025Began a one-year appointment as a member of technical staff at Thinking Machines Lab for academic year 2025–26.¹

Other documented roles

Roles the sources document that are not confirmed as current.

• Member of technical staff (academic year 2025–26), Thinking Machines Lab¹
• Research scientist, Facebook AI Research (FAIR), Meta (Facebook AI Research)¹
• Research scientist, Google DeepMind¹

Honors and recognition

• NSF CAREER Award · National Science Foundation, 2023⁵
• Sloan Research Fellowship · Alfred P. Sloan Foundation, 2022⁵
• ACL Outstanding Paper Award · Association for Computational Linguistics, 2022⁵
• Arthur Samuel Best Doctoral Thesis Award · Stanford University, 2019⁵
• Facebook Fellowship · Facebook, 2017⁵
• EMNLP Outstanding Paper Award · Conference on Empirical Methods in Natural Language Processing, 2017⁵
• ACL Outstanding Paper Award · Association for Computational Linguistics, 2016⁵
• Microsoft Research Women's Fellowship · Microsoft Research, 2013⁵

Where they work

• Princeton University 5 people in the wiki

Related people

• Tri DaoAlso at Princeton University
• Alexander WettigAlso at Princeton University
• John YangAlso at Princeton University
• Kilian LieretAlso at Princeton University
• Edward BeechingLanguage models · Training systems
• Jeff DeanTraining systems · Language models
• Colin RaffelLanguage models · Training systems
• Sam McCandlishTraining systems · Language models
• Barret ZophLanguage models · Training systems
• Luke ZettlemoyerLanguage models · Training systems