Paroma Varma
Co-Founder at Snorkel AI
United States
Summary
🤩
Rockstar🎓
Top SchoolParoma Varma is a co-founder and PhD-trained machine learning researcher with 11 years of engineering experience focused on generative models and weakly supervised workflows. She co-founded Snorkel AI after contributing to the Snorkel framework—helping build labeling functions and LSTM-based sentiment models to rapidly generate training data when labeled examples are scarce. Her background spans academia and industry, including a Stanford PhD, a software engineering stint at Meta, and research and teaching roles at UC Berkeley. Paroma combines deep theoretical knowledge with practical systems work that moves ML prototypes into usable training pipelines, and she often tackles problems where data is limited rather than abundant.
11 years of coding experience
3 years of employment as a software developer
Bachelor of Science (B.S.) Electrical Engineering and Computer Science, Bachelor of Science (B.S.) Electrical Engineering and Computer Science at University of California, BerkeleyMission San JoseDoctor of Philosophy (Ph.D.) Electrical Engineering, Doctor of Philosophy (Ph.D.) Electrical Engineering at Stanford University
Hindi, Bengali
Github Skills (7)
machine-learning10
nlp10
python10
data-science10
snorkel10
tensorflow8
data-augmentation7
Programming languages (4)
Github contributions (5)
snorkel-team/snorkel
Feb 2019 - Sep 2019
A system for quickly generating training data with weak supervision
Role in this project:
Data Scientist
Data ScientistContributions:33 commits, 97 PRs, 196 pushes in 6 months
Contributions summary:Paroma primarily contributed to the development and improvement of the Snorkel framework for generating training data with weak supervision, focusing on tasks related to sentiment analysis and image datasets. Their work involved creating and integrating labeling functions, evaluating model performance, and refining the training process, demonstrating a strong understanding of machine learning and weak supervision techniques. Key contributions included modifications to existing tutorials and the implementation of LSTM-based sentiment analysis models, demonstrating practical application of the framework.
weak-supervisionpythondata-sciencemachine-learninglabeling
paroma/paroma.github.io
May 2015 - Nov 2020
Contributions:2 PRs, 172 pushes, 4 branches in 5 years 6 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.