Durk Kingma
ML Research at Anthropic
Amsterdam, North Holland, Netherlands
Summary
🤩
Rockstar🎓
Top SchoolDurk Kingma is a machine learning researcher with 13 years of experience who helped pioneer generative modeling methods including the Variational Autoencoder (VAE) and the Adam optimizer. He holds a cum laude PhD from the University of Amsterdam and has led research teams at OpenAI, Google Brain/DeepMind, and currently Anthropic, focusing on diffusion models and large language models. Durk combines deep theoretical grounding with practical engineering—his contributions include improving training and sampling pipelines for high-profile projects like PixelCNN++. He is comfortable moving between foundational research and production-ready tooling, and was the first to receive cum laude in his department since 1985, reflecting both academic distinction and sustained impact.
13 years of coding experience
15 years of employment as a software developer
M.Sc. Computer Science (Theoretical Computer Science), M.Sc. Computer Science (Theoretical Computer Science) at Utrecht UniversityPh.D. Machine Learning, Ph.D. Machine Learning at University of AmsterdamComputer Science (visiting student), Computer Science (visiting student) at Sapienza Università di Roma
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (10)
computer-vision10
machine-learning10
tensorflow10
python10
paper10
numpy9
data-visualisation9
data-visualization9
data-visualizations9
image-processing8
Programming languages (2)
Github contributions (5)
openai/pixel-cnn
Nov 2016 - Jan 2018
Code for the paper "PixelCNN++: A PixelCNN Implementation with Discretized Logistic Mixture Likelihood and Other Modifications"
Role in this project:
ML Engineer
ML EngineerContributions:9 commits, 4 pushes in 1 year 2 months
Contributions summary:Durk primarily contributed to the training and plotting scripts within the PixelCNN++ repository. They made the scripts compatible with Python 3, fixed bugs, and improved the sample plotting functionality. Their contributions focused on data loading, model sampling, and visualization, demonstrating an understanding of the model's training and output processes.
pytorchlogisticmodificationslikelihoodpixelcnn
openai/iaf
Jun 2016 - Nov 2017
Code for reproducing key results in the paper "Improving Variational Inference with Inverse Autoregressive Flow"
Contributions:4 commits, 3 PRs, 6 pushes in 1 year 5 months
deep-learninginferenceimprovingvariationalvariational-inference
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.