VivekĀ Miglani
Research Scientist at Facebook AI
Pompano Beach, Florida, United States
Summary
š¤©
Rockstarš
Top SchoolVivek Miglani is a research scientist at Facebook AI with nine years of experience applying machine learning and signal processing to real-world problems, from speech systems to medical detection. He holds BS and MS degrees from MIT in Computer Science and Mathematics with top grades and brings prior experience across industry internships at Snapchat, SIG, and RetailMeNot as well as hands-on product work building an educational mobile app. At Facebook he focuses on advancing model interpretability, contributing to PyTorch's Captum with neuron- and layer-level conductance features and tutorials that make complex explainability techniques more accessible. His background blends rigorous academic training, practical ML engineering, and quantitative trading exposure, enabling him to bridge research and production reliably. Based in Pompano Beach, Florida, he has a track record of turning deep technical insights into usable tools and educational resources.
9 years of coding experience
2 years of employment as a software developer
Bachelor of Science (B.S.), Mathematics, 5.0/5.0, Bachelor of Science (B.S.), Mathematics, 5.0/5.0 at Massachusetts Institute of TechnologyHigh School Diploma, High School Diploma at Marjory Stoneman Douglas High School
Github Skills (4)
machine-learning10
interpretation10
python10
pytorch10
Programming languages (3)
Github contributions (5)
pytorch/captum
Aug 2019 - Dec 2022
Model interpretability and understanding for PyTorch
Role in this project:
ML Engineer & Data Scientist
ML Engineer & Data ScientistContributions:5 releases, 170 reviews, 324 commits in 3 years 4 months
Contributions summary:Vivek's commits focused on enhancing the pytorch/captum repository by adding code related to the interpretability of models and understanding the importance of layer, and neuron-level features. The commits demonstrate the addition and testing of a conductance feature, along with updates for a tutorial to cover applications in the titanic dataset. This showcases contributions in the domain of model interpretation.
pytorchinterpretable-aifeature-importanceunderstandinginterpretability
vivekmig/captum-1
Oct 2019 - Mar 2025
Model interpretability and understanding for PyTorch
Contributions:1 PR, 845 pushes, 185 branches in 5 years 5 months
pytorchunderstandinginterpretabilitydeep-learningmachine-learning
Find and Hire Top DevelopersWeāve 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.