Ofir Nachum
Researcher at OpenAI
San Francisco, California, United States
Summary
🤩
Rockstar🎓
Top SchoolOfir Nachum is a researcher and machine learning engineer with a decade of experience applying deep learning and reinforcement learning to real-world products and research problems, currently working on large language models at OpenAI in San Francisco. Previously at Google he contributed to foundational ML research—implementing neural density-ratio estimators like DualDICE and memory modules in TensorFlow—and earlier drove double-digit improvements in recommendation systems and NLP quality at Quora. He blends strong academic training from MIT (BS, MEng) with hands-on production skills in Python, C++, and TensorFlow, and a track record spanning research labs, trading firms, and industrial R&D. Less obvious: his background includes low-latency C++ systems and domain-crossing projects from rotor simulation GUIs to genetic-algorithm analyses in bioinformatics, reflecting a rare mix of theoretical depth and engineering breadth.
10 years of coding experience
10 years of employment as a software developer
Bachelor's degree, Computer Science, 5, Bachelor's degree, Computer Science, 5 at Massachusetts Institute of TechnologyAlgonquin Regional High School
English, Hebrew, German, Spanish, Chinese
Stackoverflow
21reputation
1kreached
0answers
1question
Github Skills (6)
neural-network10
machine-learning10
deep-learning10
tensorflow10
python10
ai9
Programming languages (3)
Github contributions (5)
google-research/google-research
Sep 2019 - Jan 2023
Google Research
Role in this project:
ML Engineer
ML EngineerContributions:11 commits in 3 years 4 months
Contributions summary:Ofir contributed code related to neural network approximation of density ratios using DualDICE, a machine learning technique. This involved implementing a `NeuralDualDice` class, defining placeholders, building the graph, and implementing training operations using TensorFlow. The commits demonstrate expertise in designing and implementing machine learning algorithms and working with neural networks within the context of the provided research repository.
googlemachine-learningai
tensorflow/models
Mar 2017 - Dec 2018
Models and examples built with TensorFlow
Role in this project:ML Engineer
ML EngineerContributions:3 reviews, 9 commits, 12 PRs in 1 year 9 months
Contributions summary:Ofir implemented and modified a memory module for a TensorFlow model. This module, designed for "learning to remember rare events," likely contributes to the model's ability to handle and retain specific data points. The code changes involve defining the memory class, handling key-value storage, and integrating it into the model's query process. This is evident from the code differences that focus on functions like `get_hint_pool_idxs` and `make_update_op`.
deep-learningtensorflow
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.