Miguel Martin
Senior ML Software Engineer at NVIDIA
San Francisco, California, United States
Summary
🤩
Rockstar🎓
Top SchoolMiguel Martin is a Senior ML Software Engineer with 12 years of experience building production-grade computer vision and machine learning systems across industry leaders including NVIDIA and Meta. He has deep expertise in large-scale vision datasets and models (contributions to PerceptionLM, Ego4D/Ego-Exo4D) and practical experience shipping interpretability tooling, having extended PyTorch Captum’s multimodal testing to validate feature contributions. Based in San Francisco, he blends research-grade rigor from FAIR with production reliability at NVIDIA, routinely optimizing tests and codepaths for speed and maintainability. His background spans end-to-end ML engineering—from dataset design and model evaluation to deployment and automation—grounded in an honors degree from the University of Adelaide. Notably, he balances large team projects with hands-on open-source contributions that ensure technical features are both understandable and testable.
12 years of coding experience
8 years of employment as a software developer
Bachelor of Computer Science (Honours) Computer Science, Bachelor of Computer Science (Honours) Computer Science at University of AdelaideMathematics Studies Mathematics Specialist English Communications Physics, Mathematics Studies Mathematics Specialist English Communications Physics at Edward John Eyre High School
English
Github Skills (8)
pytorch10
machine-learning10
interpretation10
pytest10
feature-selection9
python9
convolutional-neural-networks8
neural-network8
Programming languages (17)
Github contributions (5)
pytorch/captum
Sep 2019 - Dec 2020
Model interpretability and understanding for PyTorch
Role in this project:
ML Engineer & Test Automation Engineer
ML Engineer & Test Automation EngineerContributions:15 reviews, 48 commits, 23 PRs in 1 year 2 months
Contributions summary:Miguel's primary contribution was enhancing the `captum/insights` module by adding and testing multi-modal models within the `test_contribution.py` file. They implemented a `BasicMultiModal` model with image and random features, along with supporting data loading functions, to enable the testing of model interpretability features. They also added tests that verify the correctness of feature contributions. Further changes included formatting improvements and the removal of redundant code, and optimizing test speed.
pytorchinterpretable-aifeature-importanceunderstandinginterpretability
miguelmartin75/cp-problems
Jan 2019 - Jan 2025
Contributions:21 pushes, 1 branch in 6 years 1 month
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.