# Douwe Kiela

> Douwe Kiela is an AI researcher at Google DeepMind, a co-founder and former CEO of Contextual AI, and an adjunct professor in Stanford’s Symbolic Systems Program. He co-authored the 2020 retrieval-augmented generation paper and works on grounded language understanding, multimodal models, and evaluation.

- **Role:** Researcher at Google DeepMind; Adjunct Professor, Symbolic Systems Program at Stanford University
- **Organizations:** Google DeepMind, Contextual AI, Stanford University, Hugging Face, Facebook AI Research, University of Cambridge
- **From:** Amsterdam, Netherlands
- **Nationality:** Dutch-American
- **Education:** BSc in Liberal Arts & Sciences; double major in Cognitive Artificial Intelligence and Philosophy, Utrecht University, MSc in Logic, University of Amsterdam, Institute for Logic, Language and Computation, MPhil in Computer Science, University of Cambridge, PhD in Computer Science, University of Cambridge, Darwin College
- **Known for:** Co-authored Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks, accepted at NeurIPS 2020., Helped introduce Dynabench, a platform for dynamic human-and-model evaluation., First author of the 2020 Hateful Memes Challenge paper.

## Career timeline

- **2015** — Co-authored Grounding Semantics in Olfactory Perception with Luana Bulat and Stephen Clark.
- **2016** — Submitted his Cambridge doctoral dissertation and joined Facebook AI Research as a postdoctoral researcher.
- **2020** — Co-authored the RAG paper; helped introduce Dynabench and the Hateful Memes Challenge.
- **2021** — Began his Stanford adjunct appointment, as dated in his public professional history.
- **2022** — Served as Head of Research at Hugging Face.
- **2023** — Co-founded Contextual AI with Amanpreet Singh; announced a $20 million seed round on June 7.
- **2024** — Contextual AI announced an $80 million Series A led by Greycroft.
- **2026** — Co-authored a controlled study of single-agent and multi-agent reasoning with Dat Tran; joined Google DeepMind in May.

## Achievements

- Co-authored Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks, accepted at NeurIPS 2020.
- Helped introduce Dynabench, a platform for dynamic human-and-model evaluation.
- First author of the 2020 Hateful Memes Challenge paper.
- Co-founded Contextual AI, which announced $20 million in seed funding and an $80 million Series A.
- Appointed adjunct professor in Stanford’s Symbolic Systems Program.

## Latest updates

- **2026-05** — Joined Google DeepMind; Bloomberg reported that the lab licensed Contextual AI technology and hired more than 20 researchers including Kiela.
- **2026-04** — Published Single-Agent LLMs Outperform Multi-Agent Systems on Multi-Hop Reasoning Under Equal Thinking Token Budgets with Dat Tran.

## Links

- Website: https://douwekiela.github.io/
- LinkedIn: https://www.linkedin.com/in/douwekiela/
- Twitter/X: https://x.com/douwekiela
- GitHub: https://github.com/douwekiela

---

Profile page: https://yespress.io/douwe-kiela
Published by YesPress — https://yespress.io
Last updated: 2026-10-11