Martin Kim
Software Engineer at insitro
San Francisco, California, United States
Summary
🤩
Rockstar🎓
Top SchoolMartin Kim is a software engineer and UC Berkeley student bridging bioengineering and EECS with seven years of hands-on research and industry experience across academic labs and biotech. He has a strong focus on machine learning for biology, notably optimizing scvi-tools to speed inference in PyTorch for large-scale single-cell and spatial omics analysis. At insitro and prior roles at Genentech, Berkeley Lab, and university research groups he translated probabilistic models into performant, production-ready code. Based in San Francisco, he pairs deep learning expertise with practical software engineering to accelerate biological discovery in healthcare, food tech, and energy. A detail-oriented contributor, he often surfaces non-obvious performance wins by refactoring model internals and adopting inference-mode optimizations.
7 years of coding experience
3 years of employment as a software developer
Bachelor of Science - BS, Bioengineering, Electrical Engineering & Computer Science, Bachelor of Science - BS, Bioengineering, Electrical Engineering & Computer Science at UC Berkeley College of Engineering
English, Korean, French, Spanish
Github Skills (7)
pytorch10
deep-learning10
sc9
python9
variational-inference9
rna-seq9
seq9
Programming languages (6)
Github contributions (5)
scverse/scvi-tools
Sep 2022 - Jan 2023
Deep probabilistic analysis of single-cell and spatial omics data
Role in this project:
ML Engineer
ML EngineerContributions:17 releases, 408 reviews, 37 commits in 4 months
Contributions summary:Martin primarily focused on optimizing the scvi-tools library for faster inference in PyTorch, as evidenced by their commit messages detailing changes to inference modes, removal of deprecated features, and refactoring of specific model methods. They updated various model files (e.g., `_totalvi.py`, `_negative_binomial.py`, `_multivi.py`) to leverage `torch.inference_mode()` and related optimizations. Their contributions directly improved the performance of models within the scvi-tools framework, particularly in areas related to deep probabilistic analysis of single-cell data.
hierarchicaldeep-generative-modelmixture-of-expertssingle-cell-rna-seqsingle-cell
YosefLab/mrvi_archive
Oct 2022 - Feb 2024
Multi-resolution Variational Inference
Contributions:1 release, 1 review, 10 PRs in 1 year 3 months
variational-inference
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.