Vitor Guizilini
Staff Research Scientist & Team Lead (Learning From Videos) at Toyota Research Institute
San Francisco Bay Area United States
Summary
🤩
Rockstar🎓
Top SchoolVitor Guizilini is a Staff Research Scientist and team lead based in the San Francisco Bay Area who specializes in teaching robots to learn from video, with six years of industry experience focused on perception and self-supervised learning. Holding a PhD in Artificial Intelligence from the University of Sydney and advanced degrees in mechatronics from USP, he blends deep academic rigor with hands-on engineering at Toyota Research Institute, progressing from Senior Research Scientist to team lead. His open-source contributions include significant enhancements to TRI-ML’s PackNet-SfM monocular depth codebase—adding faster model variants, multi-camera support, and operational stability fixes—highlighting a pragmatic focus on both research novelty and production readiness. Comfortable leading teams and shipping systems, he brings a rare combination of robotics, vision, and systems-level optimization to autonomous perception problems.
6 years of coding experience
6 years of employment as a software developer
The University of SydneyMaster's degree Mechatronics Robotics and Automation Engineering, Master's degree Mechatronics Robotics and Automation Engineering at USP - Universidade de São Paulo
Portuguese, English, Spanish, German
Stackoverflow
31reputation
2kreached
0answers
3questions
Github Skills (15)
depth-estimation10
computer-vision10
machine-learning10
pytorch10
deep-learning10
python10
sparse-matrix6
matrix-multiplication6
tensorflow6
batch-normalization6
gpu6
gradient6
style-transfer6
dockers3
docker3
Programming languages (5)
Github contributions (5)
TRI-ML/packnet-sfm
May 2020 - Jul 2022
TRI-ML Monocular Depth Estimation Repository
Role in this project:
ML Engineer
ML EngineerContributions:1 release, 7 reviews, 8 commits in 2 years 2 months
Contributions summary:Vitor made several contributions focused on improving the PackNet-SfM project, which is a monocular depth estimation repository. Their commits include the initial release of training code, support for multi-camera loading, implementation of a velocity loss function, and the addition of PackNetSlim01, a faster model variant. They also addressed memory leaks and made inference improvements, indicating a focus on both model development and operational efficiency.
pytorchmonocular-depth-estimationdeep-learningmachine-learningdepth
TRI-ML/vidar
May 2022 - Dec 2022
Contributions:1 release, 6 reviews, 4 commits in 6 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.