David Dohan
Member Of Technical Staff at OpenAI
San Francisco Bay Area United States
Summary
🤩
Rockstar🎓
Top SchoolDavid Dohan is a Member of Technical Staff at OpenAI with 13 years of experience building and researching large-scale ML systems that augment human thought. Previously a researcher on Google Brain, he worked on LLMs for program synthesis, protein engineering, and reasoning (projects like PaLM, Minerva, ProtNLM, and Cascades). He blends research and engineering fluency—contributing to high-profile open-source efforts such as tensor2tensor and the initial QANet open-source data pipeline—while shipping production data and analytics infrastructure earlier in his career at Dropbox and Sift Science. Based in the Bay Area and trained at Princeton, he focuses on human-and-machine learning interactions and brings a pragmatic, systems-oriented approach to cutting-edge generative models.
14 years of coding experience
8 years of employment as a software developer
Bachelor of Engineering (BEng), Computer Science, Bachelor of Engineering (BEng), Computer Science at Princeton University
Github Skills (15)
neural-network10
machine-learning10
rnn-model10
convolutional-neural-networks10
deep-learning10
tensorflow10
data-pipelines10
data-pipeline10
n10
python10
ai9
api8
apim8
integrations8
api-design8
Programming languages (8)
Github contributions (5)
google-research/google-research
Oct 2018 - Jan 2023
Google Research
Role in this project:
ML Engineer
ML EngineerContributions:6 commits in 4 years 3 months
Contributions summary:David contributed to the initial open sourcing of QANet, focusing on the data pipeline. Their commits demonstrate involvement in implementing data loading and processing functionalities specific to the SQuAD dataset. This includes the modification of core data loading scripts to support the integration and utilization of the QANet model with the dataset.
googlemachine-learningai
tensorflow/tensor2tensor
Dec 2018 - Jul 2019
Library of deep learning models and datasets designed to make deep learning more accessible and accelerate ML research.
Role in this project:ML Engineer
ML EngineerContributions:7 commits in 7 months
Contributions summary:David contributed to the tensor2tensor library by implementing and modifying functions related to multi-problem support within the model, specifically focusing on weight assignments based on task IDs. They added explicit checks for task IDs within these functions. Furthermore, the user integrated support for warm-starting models from checkpoints, adding this functionality to MTF models and addressing initialization procedures within the project. They also integrated new RNN components, including the NeuralGPU, and a causal convolution layer.
deep-learningmachine-learningmachine-translationreinforcement-learningtpu
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.