Gabriel Ilharco
AI Research Scientist at Meta
Seattle, Washington, United States
Summary
🤩
Rockstar🎓
Top SchoolGabriel Ilharco is an AI research scientist with 11 years of experience bridging cutting-edge research and production ML engineering, currently at Meta after a stint at xAI. He holds a PhD from the University of Washington and advanced degrees from ITA, and has a track record at top labs including multiple roles at Google and internships at Microsoft. Gabriel contributes to prominent open-source ML tooling—most notably improving training reliability and dataset support in the widely used open_clip CLIP implementation—bringing practical robustness to research models. He combines research depth (AI residency and research internships) with hands-on systems work, from metric and data-pipeline fixes to release management, and often focuses on making research tooling more usable and testable. Colleagues would describe him as someone who moves models toward reliable, reproducible deployment without losing sight of novel research directions.
11 years of coding experience
2 years of employment as a software developer
Doctor of Philosophy - PhD, Doctor of Philosophy - PhD at University of WashingtonMaster of Science (M.S.), Master of Science (M.S.) at Instituto Tecnológico de Aeronáutica - ITA
English, Spanish, Portuguese
Stackoverflow
1,669reputation
167kreached
38answers
11questions
Github Skills (14)
computer-vision10
pytorch10
deep-learning10
python10
pre-trained-model9
data-engineering8
wandb8
webdataset8
matlab6
javascript6
jquery6
pascal6
python-requests6
java6
Programming languages (7)
Github contributions (5)
mlfoundations/open_clip
Jul 2021 - Jan 2023
An open source implementation of CLIP.
Role in this project:
ML Engineer
ML EngineerContributions:10 reviews, 21 commits, 61 PRs in 1 year 5 months
Contributions summary:Gabriel primarily contributed to the development and maintenance of the open_clip repository, an open-source implementation of CLIP. Their work involved fixing metrics and data filtering processes within the training pipeline. They also focused on improving the usability and functionality of the project by making the wandb project name configurable and setting the wandb id. Furthermore, the user released multiple versions of the project and added support for dataset mixtures and DataComp models, and made improvements to the testing of the project.
pytorchlanguage-modelzero-shot-classificationdeep-learningclip
mlfoundations/patching
Jun 2022 - Oct 2022
Contributions:12 commits, 10 pushes, 1 branch in 3 months
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.