Allan J
Member Of Technical Staff at OpenAI
San Francisco, California, United States
Summary
🤩
Rockstar🎓
Top SchoolAllan J is a Member of Technical Staff at OpenAI with 11 years of experience building research-driven software systems that bridge cutting-edge ML research and production-grade tooling. He holds a PhD from UC Berkeley and a BS from Princeton, and has research and engineering stints at DeepMind, Google Brain, and Facebook AI Research where he focused on scalable ML environments and tooling. Allan contributes to notable open-source projects—improving core serializers and remote-learner support for facebookresearch/CommAI-env and adding PyTorch visualization and image handling improvements to fossasia/visdom—demonstrating strength across back-end and full-stack domains. He combines deep research exposure with pragmatic engineering: shipping serializer and socket-based remote learner features as well as UI and tensor-visualization fixes. Based in San Francisco, he often operates at the intersection of live-system engineering and research prototypes, turning experimental ideas into robust developer-facing tools. An understated thread through his work is attention to developer experience, from clearer serialization to higher-quality visual artifacts.
11 years of coding experience
4 years of employment as a software developer
Doctor of Philosophy - PhD, Doctor of Philosophy - PhD at University of California, BerkeleyBachelor of Science, Bachelor of Science at Princeton University
Github Skills (14)
pytorch10
data-serialization10
visualization10
zeromq10
visualizations10
python10
back-end-development10
serialization10
react9
javascript9
socket9
numpy8
image-processing8
web-development7
Programming languages (10)
Github contributions (5)
facebookresearch/CommAI-env
Sep 2016 - Jul 2017
A platform for developing AI systems as described in A Roadmap towards Machine Intelligence - http://arxiv.org/abs/1511.08130
Role in this project:
Back-end Developer
Back-end DeveloperContributions:42 commits, 2 PRs, 2 pushes in 10 months
Contributions summary:Allan primarily contributed to the core functionalities of the environment, focusing on improving the internal serializer and implementing a non-interactive viewing mode. These changes involved modifications to the `serializer.py` file, enhancing readability and functionality. Additionally, the user worked on implementing remote learner support, allowing learners to be expressed as arbitrary binaries and communicating with the environment via sockets.
ai
fossasia/visdom
Mar 2017 - Jan 2018
Tool for real-time visualization, monitoring and collaborative analysis of AI/ML experiments and live data. Supports Python, PyTorch/Torch, NumPy, TensorFlow/Keras https://visdom.dev
Role in this project:Full-stack Developer
Full-stack DeveloperContributions:10 commits, 14 PRs, 9 pushes in 9 months
Contributions summary:Allan primarily contributed to the Visdom project by implementing features and fixing bugs related to its user interface and core functionality. They introduced PyTorch support, enabling the visualization of PyTorch tensors. The user also addressed formatting and display issues within the text panes and improved the image handling process. Furthermore, they updated the project dependencies and improved the image quality for temporary saving.
aikerasmachine-learningmonitoringnumpy
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.