Heitor Guimarães
Machine Learning Researcher at RBC Borealis
Montreal, Quebec, Canada
Summary
👤
Senior🎓
Top SchoolHeitor Guimarães is a Machine Learning Researcher and final-year PhD candidate at INRS specializing in efficient deep learning, OOD robustness, and AI security, with applied work in speech and bioacoustics. With 11 years of industry and research experience, he has interned as a research scientist at Adobe and Meta and now contributes to RBC Borealis on production-scale ML. He blends classical tabular-data expertise from banking and O&G with state-of-the-art work on representation learning, model compression, and diffusion/multimodal models. An active open-source contributor, he has improved SpeechBrain (notably multichannel SincConv and Whisper integration), reflecting his focus on practical toolchains for speech research. Heitor enjoys translating complex research into systems that impact millions and communicating results to both technical and non-technical audiences.
11 years of coding experience
12 years of employment as a software developer
Especialização, Data Science, Especialização, Data Science at Instituto Tecnológico de Aeronáutica - ITAMaster of Science, Artificial Intelligence, Master of Science, Artificial Intelligence at USP - Universidade de São PauloBachelor's degree, Computer Engineering, Bachelor's degree, Computer Engineering at Federal University of Rio de JaneiroDoctor of Philosophy - PhD, Machine Learning, Speech and Acoustic Signal Processing, Doctor of Philosophy - PhD, Machine Learning, Speech and Acoustic Signal Processing at INRS - Institut national de la recherche scientifique
Portuguese, English, French
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (12)
transformer-models10
huggingface-transformers10
pytorch10
speech-recognition10
deep-learning10
speech-processing10
asr9
python9
audio-processing9
faster-rcnn8
mask-rcnn8
nlp7
Programming languages (5)
Github contributions (5)
speechbrain/speechbrain
Feb 2022 - Dec 2022
A PyTorch-based Speech Toolkit
Role in this project:
ML Engineer
ML EngineerContributions:3 reviews, 5 commits, 4 PRs in 10 months
Contributions summary:Heitor contributed to the SpeechBrain repository by implementing and modifying various aspects of the speech processing toolkit. Their work included adding support for multichannel SincConv, fixing model evaluation issues, enabling the retrieval of hidden states from the Hugging Face Whisper model, and updating the Whisper model to align with the wav2vec2 class. Furthermore, the user improved documentation and addressed hypermixing-related issues within the transformer models.
voice-recognitionasrspeech-recognitionspeech-separationspeaker-verification
Hguimaraes/learnfy
Feb 2017 - Jul 2017
Contributions:36 commits, 1 push in 5 months
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.