Education
-
PhD Student in Robotics Engineering - Present
-
DLS IIT - Genova, Italia
-
-
Master's degree in Mechanical Engineering - 2023
-
Pontifical Catholic University - Rio de Janeiro, Brazil
-
-
Bachelor's degree in Mechanical Engineering - 2020
-
Pontifical Catholic University - Rio de Janeiro, Brazil
-
-
Exchange Student in Mechatronic Engineering - 2018
-
Ernst-Abbe Hochschule - Jena, Germany
-
Gabriel Fischer Abati
Hello there! I am a research engineer focusing in robotics and deep learning. During my Bachelor's I was a member of the robotics team, Riobotz, where I learned about practical mechanical/electronic design and construction, as well as, programming autonomous systems, control and dynamics. I did my Master's thesis about mobile robotics, SLAM and deep learning tecniques at the Robotics Lab (LabRob) from PUC-Rio University. In 2022, I worked at fu2re Smart Solutions Company with deep learning projects for the energy sector. Currently, I am a PhD Student at the dynamic legged systems lab in Istituto Italiano di Tecnologia (IIT) working with robot state estimation, deep learning and autonomous navigation. You will find all of my projects and research papers on this website. Feel free to browse around and contact me.
Interests
-
Deep Learning
-
Computer Vision
-
Robotics
-
Dynamics and Control
Awards
-
Second place in CTDR - 2023
-
Second place in Shell LiveWire - 2018
-
Most Destructive Robot Award in 113 Kg Combat Robots - 2016
-
Gold medal in heavy weight (100 Kg) Combat Robots - 2015
Rio de Janeiro, Brazil
Battlebots - United States
RoboGames - United States
Bahia, Brazil
Professional Experience
-Artificial Intelligence Researcher
fu2re Smart Solutions - Rio de Janeiro, Brazil
-
Industrial Gas Canister Identification: Developed and implemented a system to identify gas canister identification tags. The system runs in a NVIDIA Jetson board with trained deep learning models to inference on canisters images.
2022 - Present
-
Python
-
Deep Learning
-
Docker
-
Computer Vision
-
Edge Computing
-Computer Vision Engineer
2019 - 2022
Cyberlabs - Rio de Janeiro, Brazil
-
Bio-metric authentication API: Developed and implemented a facial recognition model concerning with ethnicity bias within the Brazilian population. Proposed a new visual anti-spoof method based on computer vision techniques to make facial recognition model safer against spoof attacks. Studied multi-task learning models to improve robustness from face recognition model. Combined face and speaker recognition models to enhance bio-metrics scanning.
-
Airport wait time measurement: Project to measure the wait time of inspection procedure at Galeao airport (Rio de Janeiro, Brazil). I developed a robust multi-camera system that detects a person at the entrance and exit gate and matches them using Deep Learning methods.
-
Python
-
Pytorch
-
Docker
-
Torchserver
-
Deep Learning
-
Computer Vision
-
C++
-
Python
-
Pytorch
- Nvidia DeepStream
-
Docker
- Deep Learning
Skills
Programming Languages
-
C/C++
-
Python
-
Matlab/Simulink
-
CUDA SDK
Robotics
- ROS
-
Embedded systems
- Control techniques (PID,LQR,LQG,...)
Deep Learning Frameworks
- Pytorch
-
OpenCV
- DeepStream
- Tensorflow
Machine Learning Frameworks
- Pandas
- Numpy
- Scikit-learn
Languages
-
English
-
Cambridge English Level Test - EFCELT-B2
-
Excellent level for writing,reading,speaking and comprehension
-
-
German
-
Deutsches Sprachdiplom der Kulturministerkonferenz Zweit Stufe - C1
-
Excellent level for writing,reading,speaking and comprehension
-
-
Portuguese
-
Native language
-
Mechanical Design
- SolidWorks
- Ansys
PCB Design
- KiCad
Backend
- Git
-
Linux
- Docker
- Torchserver