John Arevalo
Principal Machine Learning Scientist at Genentech
South San Francisco, California, United States
Summary
👤
Senior🎓
Top SchoolJohn Arevalo is a Principal Machine Learning Scientist with 12 years of experience building predictive systems at the intersection of deep learning, medical imaging, computer vision, NLP and multimodal learning. He holds a PhD in Computer Science and has a track record of influential academic work—including ICLR-published multimodal attention research and highly cited CNN applications for mammography—that has been adopted across architectures like CNNs, RNNs, GANs and Transformers. After foundational research roles across academia and industry, he moved from postdoctoral work at the Broad Institute to applied ML leadership at AI Fund and now leads ML science at Genentech in South San Francisco. An active open-source maintainer early in the Theano era, he improved core libraries like blocks and pylearn2 to enhance usability and backwards compatibility, reflecting a pragmatic focus on reliable tooling as well as novel models.
12 years of coding experience
17 years of employment as a software developer
Doctor of Philosophy - PhD Computer Science, Doctor of Philosophy - PhD Computer Science at Universidad Nacional de ColombiaBachelor of Engineering - BE Computer Science, Bachelor of Engineering - BE Computer Science at Universidad Distrital Francisco José de Caldas
English, Spanish
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (13)
data-preprocessing10
machine-learning10
tensorflow10
python10
theano10
numpy9
testing9
machine-learning-models9
neural-network9
web-framework9
deep-learning8
algorithms8
documentation7
Programming languages (9)
Github contributions (5)
mila-iqia/blocks
Jun 2015 - Feb 2017
A Theano framework for building and training neural networks
Role in this project:
ML Engineer
ML EngineerContributions:6 commits, 6 PRs, 4 comments in 1 year 8 months
Contributions summary:John's commits primarily focus on improving the functionality and documentation within the `blocks` framework. They addressed syntax errors in documentation, expanded the types supported by lookup indexing, and refined hyperparameters within the algorithms. The user updated deprecated parameters in the convolution and pooling modules. These changes suggest a focus on maintaining and improving the usability and correctness of the core machine-learning library.
pytorchdeep-learningtheanoneural-networksmachine-learning
lisa-lab/pylearn2
Mar 2014 - Jul 2014
Warning: This project does not have any current developer. See bellow.
Role in this project:ML Engineer
ML EngineerContributions:6 commits, 1 PR, 1 comment in 3 months
Contributions summary:John primarily contributed to the pylearn2 project by addressing compatibility issues, specifically adding backward-compatibility with Python 2.6 format syntax. The user modified several files related to data loading, preprocessing, and model definitions, including `svhn.py` and `svhn_preprocessing.py`, demonstrating a focus on data handling and preparation within the context of machine learning. Furthermore, the user made changes to the testing framework, enabling and correcting tests for the MLP model.
javascripttypescript
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.