Career Profile
Olá! My name is Angelo Garangau Menezes. I’m currently a Ph.D. candidate at the University of São Paulo, where I research machine learning, focusing on continual learning applied to computer vision tasks. Previously, I worked with R&D of embedded systems (software and hardware) in Brazil, Canada, and Germany.
Experiences
- Worked on improving search vertical features on the search engine for more exploratory user journeys.
- Investigated how self-supervised representations learned from videos behave when applied to solve different downstream tasks.
- Provided expertise for creating object detection pipelines with better generalization.
- Taught the course “Advanced Topics in Intelligent Systems: Deep Learning” in the “Big Data and Intelligent Systems’’ specialization program.
- Managed a group of two engineering interns and a technician to develop state-of-the-art electronic instruments for various applications such as industrial automation and domotics. Main activities included prototyping, schematics and board design, selecting electronic components and microcontroller programming (C and C++).
- Elaborated strategies to solve technical and commercial issues related to engineering projects.
- Documented all technical aspects of the developed projects.
- Designed computer vision pipelines with low computational resources (Raspberry Pi series).
- Programmed PLCs and HMIs for automation systems.
- Developed the electronics and documentation behind a module for voltage conversion based on motor inclination in less than 3 months.
- Developed an internal library in C++ for applying 1D Kalman filtering in angle/acceleration data.
- Programmed the voltage conversion module MCUs (Atmega and PIC) in C and C++.
- Developed the electronics and software behind a smart sensor system for 3D automatic acceleration/vibration measurement with high bandwidth and measurement ranges for the water treatment plant of Thunder Bay.
- Programmed the smart sensor MCU (Atmega and PIC) in C and C++.
- Developed a user interface in C# for the smart sensor system.
Projects
3rd place in the Continual Instance Detection Competition
- CLVision Workshop at CVPR 2022
Automatic Attendance Management via Face Recognition
- An API for automating the process of attendance registration in a classroom using facial recognition. (In Portuguese)
Face Super-Resolution
- Implementation of different face super-resolution algorithms in Pytorch for improving low-resolution face recognition.
DICOM Labeling Tool
- Simple web-based app for labeling DICOM files written in Python.
3D Face Reconstruction
- An implementation of a 3D face reconstruction model that extracts depth directly from low-resolution images.
Image Processing
- Some image processing algorithms written from scratch for image enhancement, restoration, and segmentation.
Selected Publications
Continual object detection: a review of definitions, strategies, and challenges
Neural Networks (2023)
Automatic Attendance Management System based on Deep One-Shot Learning
27th International Conference on Systems, Signals and Image Processing (IWSSIP 2020)
Supervised Learning in the Context of Educational Data Mining to Avoid University Students Dropout
18th IEEE International Conference on Advanced Learning Technologies (ICALT 2019)
Automatic speech recognition using Support Vector Machine and Particle Swarm Optimization
IEEE Symposium Series on Computational Intelligence (SSCI 2016)