Diederik P. Kingma
Also known as Durk Kingma
Machine learning researcher at Anthropic; co-created the VAE and the Adam optimizer.
At a glance
Also on this page: Key facts · Links · Sources · Related pages
- Diederik "Durk" Kingma is a Dutch computer scientist working on machine learning and generative models.
- His PhD work at the University of Amsterdam produced the Variational Autoencoder and, with Jimmy Ba, the Adam optimizer.
- He was part of the founding team of OpenAI in 2015.
- He worked at Google Brain and DeepMind from 2018 to 2024, then joined Anthropic.
Key facts
- Work
- Research on large-scale machine learning at Anthropic (from 2024)1
- Earlier
- OpenAI founding team, led the Algorithms team (2015 to 2018); Google Brain / DeepMind (2018 to 2024)1
- Created
- Variational Autoencoder (VAE), Adam optimizer, Glow, Variational Diffusion Models1
- PhD
- University of Amsterdam (2017), advised by Max Welling1
- Honours
- ICLR Test of Time Award for the VAE paper (2024) and the Adam paper (2025)1
- Co-founded
- Advanza (2010), acquired in 20161
Roles
Part of
- AnthropicOrganization
research on large-scale machine learning · since 2024 source on Anthropic
Before
Part of until 2024
Google DeepMind (research scientist at Google Brain / DeepMind) source on Google DeepMind
Part of until 2018
OpenAI (founding team; research scientist, led the Algorithms team) source on OpenAI
Links
Sources
Last checked against its sources today
Is something on this page wrong or out of date? Tell us
The line under a date fills as time passes: empty when new, then one equal share each for a day, a week, a month and a year. For episodes and other items it is full at 2 years; for "last checked", at a year. A striped line means the date is known only to the month or year.