# Durk Kingma

Research and product · Researcher at Anthropic · joined 2024

Co-created the Adam optimizer and the variational autoencoder, two of the most used ideas in deep learning.

Full profile: https://www.autumn.ai/resources/anthropic/people/durk-kingma

**At Anthropic:**
- Oct 2024 – present: ML Research
Sources: [1](https://www.linkedin.com/in/durk-kingma-58b3564)

**Career before Anthropic:**
- Jul 2018 – Oct 2024: Google, Research Scientist, Google Brain / DeepMind
- Dec 2015 – Jul 2018: OpenAI, Research Scientist and Team Lead
- 2012–2013: Courant Institute of Mathematical Sciences, Research Scholar
- May 2009 – Dec 2009: Courant Institute of Mathematical Sciences, Jr. Research Scientist
- Feb 2009 – 2012: Advanza, Co-Founder
Sources: [1](https://www.linkedin.com/in/durk-kingma-58b3564)

**Education:**
- Ph.D., Machine Learning, University of Amsterdam, 2013-2017
- M.Sc., Computer Science (Theoretical Computer Science), Universiteit Utrecht, 2007-2009
- Computer Science (visiting student), Sapienza Università di Roma, 2006-2007
- B.Sc., Computer Science, Universiteit Utrecht, 2002-2007
Sources: [1](https://www.linkedin.com/in/durk-kingma-58b3564)

**Lesser-known facts:**
- Goes by the Frisian nickname Durk; his formal name is Diederik. [1](https://dpkingma.com/)
- Won ICLR Test of Time awards for the VAE (2024) and Adam (2025) papers. [1](https://dpkingma.com/)
- Was an early OpenAI research scientist before Google Brain. [1](https://www.linkedin.com/in/durk-kingma-58b3564)
- Co-founded Advanza and was its technical lead from 2010 to 2012; the company had a successful exit in 2016. [1](https://dpkingma.com/)
- Worked as a junior research scientist in Yann LeCun's lab at NYU in 2009 and 2012. [1](https://dpkingma.com/)
- Earned his PhD cum laude, the first such distinction in the University of Amsterdam's computer science department in 30 years. [1](https://dpkingma.com/)
- Led OpenAI's Algorithms team, focused on basic research, as part of its founding team from 2015 to 2018. [1](https://dpkingma.com/)
- The Adam paper alone has more than 266,000 citations on Google Scholar. [1](https://scholar.google.com/citations?user=yyIoQu4AAAAJ&hl=en)
- Urged people not to support the AI anti-regulation lobby, arguing some AI uses are net negative unless well regulated. [1](https://x.com/dpkingma/status/1973269967484379291)

**Interviews and talks:**
- [ICLR 2025 Test of Time talk on Adam, with Jimmy Ba (2025)](https://x.com/dpkingma/status/1915700374511763950)

**Known for:**
- Variational Autoencoder (VAE)
- Adam optimizer
- Glow
- Variational Diffusion Models
Sources: [1](https://dpkingma.com/)

Researched with Autumn (https://www.autumn.ai) from public sources. Updated 2026-10-18. Everyone at Anthropic: https://www.autumn.ai/resources/anthropic/people