Eder Santana
Sr. Applied Scientist at Twitch
Mountain View, California, United States
Summary
🤩
Rockstar🎓
Top SchoolEder Santana is a senior applied scientist based in Mountain View with 12 years of experience building and shipping machine learning systems at companies including Twitch, Toyota Research Institute, comma.ai, and Apple. He holds a PhD from the University of Florida and combines deep research experience with hands-on engineering, from self-driving stacks to production ML at scale. An active open-source contributor to Keras, he implemented masking for sequence loss computations and helped optimize core components, reflecting a focus on practical, robust deep learning infrastructure. Eder also improves developer experience through clear documentation and tooling—evident in contributions to Theano-based projects and TensorFlow subpixel super-resolution code. He prefers portfolios that show tangible projects (with pictures), and maintains a public projects showcase that reveals a pragmatic, visual approach to evaluating technical work.
12 years of coding experience
13 years of employment as a software developer
Master's degree Electrical and Electronics Engineering, Master's degree Electrical and Electronics Engineering at Universidade Federal do MaranhãoDoctor of Philosophy (Ph.D.), Doctor of Philosophy (Ph.D.) at University of Florida
English, Portuguese, memes
Github Skills (17)
convolutional-neural-networks10
python10
machine-learning10
super-resolution10
keras10
deep-learning10
tensorflow10
neural-network10
computer-vision10
documentation10
bokeh9
optimizers9
optimizer9
numpy8
data-science7
Programming languages (10)
Github contributions (5)
atriumlts/subpixel
Aug 2016 - Sep 2017
subpixel: A subpixel convnet for super resolution with Tensorflow
Role in this project:
ML Engineer
ML EngineerContributions:48 commits, 8 PRs, 21 pushes in 1 year 1 month
Contributions summary:Eder primarily contributed to the development of a subpixel convolutional neural network for super-resolution, modifying and implementing core functionalities of the model. Their work involved refactoring existing code, including changes to the `ponynet` (renamed to `subpixel`) function, and integrating it with the main model architecture within `model.py`. They also worked on integrating the subpixel network with existing code and a notebook, likely for testing and visualization.
convnetresolutionsuper-resolutiontensorflowconvolutional-neural-networks
keras-team/keras
Jul 2015 - Sep 2016
Deep Learning for humans
Role in this project:ML Engineer
ML EngineerContributions:65 commits, 52 PRs, 649 comments in 1 year 2 months
Contributions summary:Eder primarily contributed to the implementation of masking functionality within the Keras framework, focusing on loss functions for sequence data. They added masking capabilities to the `Sequential` and `Graph` models, ensuring correct loss calculations for variable-length sequences. Their work included correcting and refining these masking features, and creating tests to validate their implementation. The user also contributed to the optimization of the Keras library by making improvements to its optimizers.
pythondata-sciencedeep-learningneural-networksmachine-learning
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.