Skip to content
View oalvarobraz's full-sized avatar
  • Universidade Federal de Ouro Preto (UFOP)
  • João Monlevade, MG, Brasil
  • 07:24 (UTC -03:00)
  • LinkedIn in/alvaro-braz-cunha

Block or report oalvarobraz

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don't include any personal information such as legal names or email addresses. Markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
oalvarobraz/README.md


🧠 About Me

I am a Machine Learning Engineer specializing in Deep Learning and Computer Vision. My work focuses on building end-to-end AI solutions, from training custom neural architectures (Transformers, CNNs, U-Nets) to deploying them into production environments.

  • 🔭 Currently focused on Deep Learning with PyTorch and Cloud Infrastructure (AWS).
  • ⚡ Expertise in Custom Neural Architectures and Multi-Task Learning.
  • 🌱 Building scalable AI pipelines with Docker and AWS Lambda.
  • 🎯 Goal: Leveraging AI to solve real-world problems with high-performance engineering.

🛠️ Tech Stack & Tools

Python PyTorch AWS FastAPI Pandas Numpy ScikitLearn



Git Docker Linux PostgreSQL


📊 GitHub Stats

Snake animation

Pinned Loading

  1. blind-assist-ai blind-assist-ai Public

    A Neural Image Captioning system built from scratch with PyTorch. Features a ResNet + Transformer architecture, Beam Search inference, and Text-to-Speech (TTS) for accessibility.

    Python

  2. ClassicMLAlgorithms ClassicMLAlgorithms Public

    Implementação do zero (from scratch) de algoritmos clássicos de Machine Learning usando apenas Python e NumPy. Foco na compreensão matemática e lógica por trás de modelos como SVM, Random Forest, K…

    Jupyter Notebook

  3. eco-sort-ai eco-sort-ai Public

    Sistema End-to-End de classificação de reciclagem usando Deep Learning (ResNet50), com API em FastAPI e interface web em Streamlit.

    Jupyter Notebook

  4. NeuroSeg-PyTorch NeuroSeg-PyTorch Public

    Segmentação de Tumores Cerebrais com U-Net e PyTorch. Pipeline completo com Data Augmentation avançado, treino com precisão mista e Dice Score de ~0.85.

    Jupyter Notebook

  5. ppe-detector-yolo ppe-detector-yolo Public

    Safety-Vision: Real-time PPE detection system using YOLOv8 and PyTorch. Monitors helmets and vests usage in construction sites.

    Jupyter Notebook

  6. ThumbNet ThumbNet Public

    Sistema de Visão Computacional com Multi-Task Learning (ResNet18) que analisa thumbnails do YouTube para classificar categorias e prever visualizações simultaneamente.

    Jupyter Notebook