João Santos
Senior Research Scientist at NVIDIA
Vancouver, British Columbia, Canada
Summary
👤
Senior🎓
Top SchoolJoão Santos is a Senior Research Scientist with 16 years of experience applying AI to speech, audio and language processing, currently driving research at NVIDIA from Vancouver. His work spans speech enhancement, synthesis and objective quality/intelligibility metrics, including early end-to-end synthesis work on the Char2Wav project at Mila. He combines deep academic training (PhD in Telecommunications) with hands-on engineering—from firmware and DSP for hearing aids to contributions in major open-source deep learning projects like Keras and Chainer. Known for bridging theory and production, he designs algorithms that are both perceptually aware and deployment-minded, with a specialty in metrics tailored for hearing-impaired listeners.
16 years of coding experience
8 years of employment as a software developer
Engineer, Electrical Engineering, Engineer, Electrical Engineering at Universidade Federal de Santa CatarinaDoctor of Philosophy (PhD), Telecommunications, Doctor of Philosophy (PhD), Telecommunications at Université du Québec - Institut national de la recherche scientifiqueTechnician, Electronics, Technician, Electronics at Centro Federal de Educação Tecnológica de Santa Catarina
English, Portuguese, French
Github Skills (15)
neural-network10
cuda10
keras10
machine-learning10
deep-learning10
tensorflow10
python10
numpy10
deeplearning-ai9
cublas9
hdf9
data-science8
cudnn7
c-language6
cprogramming-language6
Programming languages (18)
Github contributions (5)
keras-team/keras
Apr 2015 - Jan 2016
Deep Learning for humans
Role in this project:
ML Engineer
ML EngineerContributions:26 commits, 19 PRs, 121 comments in 9 months
Contributions summary:João made several contributions related to the Keras deep learning library. These contributions include fixing bugs, such as allowing the use of classes that emulate the ndarray API. They also updated the code to support HDF5 datasets and incorporated changes from upstream master. Moreover, the user added functionality for saving and loading model weights to/from HDF5 files.
pythondata-sciencedeep-learningneural-networksmachine-learning
chainer/chainer
Jul 2015 - Aug 2015
A flexible framework of neural networks for deep learning
Role in this project:Back-end Developer
Back-end DeveloperContributions:6 commits, 2 PRs, 3 comments in 16 days
Contributions summary:João contributed to the `chainer/chainer` repository by implementing and modifying mathematical functions. They added and modified functions related to `sin` and `cos` calculations. The user fixed issues in existing code, including superclass references and numpy namespace usage. Additionally, they addressed code style issues and input indexing.
cudapythonmxnetcaffe2flexible-framework
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.