Jeffrey Wu
Member Of Technical Staff at Anthropic
Mountain View, California, United States
Summary
🤩
Rockstar🎓
Top SchoolJeffrey Wu is a research-oriented software engineer with 14 years of experience building and deploying large-scale ML and distributed systems, currently a Member of Technical Staff at Anthropic in Mountain View. He worked on early language-model research and RLHF at OpenAI, contributing to GPT-2 core functionality and production deployment (including nucleus sampling and multi-model Dockerized builds) and later led projects on scalable oversight, generalization, and interpretability. His background spans deep learning infrastructure at Google, founding engineering on cloud container tooling, and practical MLOps work that moved research models into production and cloud storage. Comfortable across ML, security, and systems, he combines rigorous MIT training in CS/math with a knack for shipping reproducible, deployment-ready research systems.
14 years of coding experience
12 years of employment as a software developer
Master's degree Computer Science, Master's degree Computer Science at Massachusetts Institute of Technology
Github Skills (26)
continuous-deployment10
docker10
python10
machine-learning10
dockers10
ml-deployment10
trainings10
tensorflow10
gpt10
nlp10
modeling10
react10
javascript9
typescript9
cloud-infrastructure9
Programming languages (18)
Github contributions (5)
openai/gpt-2
Feb 2019 - Dec 2020
Code for the paper "Language Models are Unsupervised Multitask Learners"
Role in this project:
ML Engineer
ML EngineerContributions:35 commits, 25 PRs, 67 pushes in 1 year 10 months
Contributions summary:Jeffrey contributed to the core functionality of the GPT-2 model by adding and modifying scripts for model interaction and sample generation. They updated the download script and Dockerfiles to support different model sizes and deployment environments. The user implemented nucleus sampling, a key technique for improving the quality of text generation. Furthermore, they fixed bugs and updated model-related code, including changes to the download process and improvements to the user interface.
nlpunsupervisedlearnersmachine-learninglanguage-models
nshepperd/gpt-2
Feb 2019 - Dec 2020
Code for the paper "Language Models are Unsupervised Multitask Learners"
Role in this project:MLOps Engineer
MLOps EngineerContributions:26 commits in 1 year 9 months
Contributions summary:Jeffrey's contributions center on the deployment and configuration of the GPT-2 model. They updated the download script, optimized the build process with Dockerfiles, and integrated multiple model versions, including 345M, 774M, and 1558M models. Further, the user incorporated nucleus sampling, enhancing model generation capabilities, and migrated the model's storage location to Azure Blob Storage.
nlpunsupervisedlearnersmachine-learninglanguage-models
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.