Kyle Gao
Member Of Technical Staff at Anthropic
New York, New York, United States
Summary
🤩
Rockstar🎓
Top SchoolKyle Gao is a seasoned software and machine learning engineer with 13 years of experience building production-scale AI and data platforms, currently a Member of Technical Staff at Anthropic in New York. He has led engineering and ML teams across startups and enterprises—as CTO/co‑founder of Skipper and an uber tech lead at Abnormal Security—delivering event-driven systems for security and drug-discovery ML services at BenevolentAI. His background combines deep research pedigree (CMU, IBM research, IJCAI publication) with hands-on systems work, including maintaining IBM's open-source pytorch-seq2seq framework and improving checkpointing for robust model training. Kyle is comfortable bridging product, research, and infrastructure: hiring and managing distributed teams while rolling up his sleeves on backend and ML engineering. He brings a track record of turning complex scientific models into reliable services and has experience optimizing real-time detection pipelines and resumable ML training workflows.
13 years of coding experience
11 years of employment as a software developer
YumingBachelor's degree Computer Science, Bachelor's degree Computer Science at Beihang UniversityMaster of Language Technologies Information Retrieval, Master of Language Technologies Information Retrieval at Carnegie Mellon UniversityExchange Student Computer Science, Exchange Student Computer Science at National University of Singapore
English, Chinese
Github Skills (8)
pytorch10
checkpoint10
back-end-development10
python10
checkpointing10
serialization9
machine-learning9
data-serialization9
Programming languages (4)
Github contributions (5)
IBM/pytorch-seq2seq
Jun 2017 - May 2018
An open source framework for seq2seq models in PyTorch.
Role in this project:
Back-end Developer
Back-end DeveloperContributions:1 release, 132 commits, 86 PRs in 10 months
Contributions summary:Kyle primarily focused on code related to checkpoint management within the PyTorch seq2seq framework. They made several changes to the `seq2seq/util/checkpoint.py` file, which involved saving and loading model parameters, optimizer states, and vocabularies. The user also made related changes to test files and Supervised Trainer to ensure that model training can be resumed correctly after a checkpoint. These changes appear to be focused on improving model training and checkpointing functionality.
deep-learningpytorchvisual-recognitionseq2seq
kylegao91/aoc-2020
Dec 2020 - Dec 2020
Contributions:23 pushes, 1 branch in 21 days
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.