Hello! I am a junior full stack engineer with over four years of experience in computational agronomy. I started my career as an intern and currently work as a full stack engineer at the same company. My professional journey has been driven by the desire to unleash the power of technology to simplify people's lives and solve real-world problems.

In recent years, I’ve honed my skills in deep learning and computer vision, exploring convolutional neural networks (CNNs), transfer learning, and image processing for various projects. I’ve also delved into MLOps practices, orchestrating pipelines with Amazon SageMaker, handling data versioning, and applying experiment tracking. My focus is to blend a solid theoretical background with hands-on development to deliver impactful AI-driven solutions.

Currently, I'm studying Transformers (Transformers) and LLMs (Large Language Models) and their applications in computer vision.

• Python: Solid four years of experience in software development and image processing.
• Software Engineering: Best practices in code development, CI/CD, testing, packaging and maintenance.
• Computer Vision / Deep Learning: Intermediate knowledge of CNNs, data augmentation, transfer learning, and key metrics (F1, IoU, mAP, etc.).
• Transformers: Intermediate knowledge of Transformers, LLMs, and their applications in natural language processing. Have used the HuggingFace Transformers library to build and train models like OwnV2, SAM2, CLIP, and Visual Transformer.
• Machine Learning: Practical experience building and deploying ML models.
• MLOps: Experience using SageMaker, pipelines, and experiment tracking (TensorBoard, MLflow).
• AWS: Working knowledge (S3, ECR, and SageMaker); pursuing more advanced certifications.

• OpenCV (cv2): Image processing, homography, segmentation, and computer vision.
• NumPy: Efficient numerical computing, data manipulation, and analysis.
• Matplotlib: Data visualization.
• Scikit-Learn: Machine learning tools.
• PyTorch: Building and training neural networks.

• Global Platform Integration: Implementation of services and integration of a global platform, developing skills in English communication and image processing with global stakeholders.
• MLOps Pipeline: Development of pipelines for classification and object detection neural networks using SageMaker, S3, and ECR.
• Homography Project: Utilization of image processing and homography to map corresponding points in different images to auto label images from different sources.
• Cloud Segmentation: Application of image processing techniques to segment clouds in satellite images.
• Auto Label: Application of software engineering to mine data from different sources and auto classify images to create a classification dataset. After that, used OwnV2 to auto detect objects in images and create a detection dataset with a zero-shot approach.
• Vision Transformer: Development of a visual transformer pipeline in Sagemaker using the HuggingFace Transformers library to classify images.
• Bounding Box Generation: Used OwnV2 to generate bounding boxes in datasets where I knew only the image content with 98% accuracy.

• AWS Certification: Aim to obtain Machine Learning certification.
• International Career: Development of an international career plan.
• Knowledge Sharing: Interest in sharing acquired knowledge with the team.
• Continuous Learning: Pursue knowledge in new technologies and tools.
• Agile Methodologies: Interest in Agile methodologies and their application in software development.
• Problem Solving: Interest in problem solving and the application of software engineering to solve real-world problems.
• Fast Learning: Fast and self-taught learner.
• Multidisciplinary: Interest in interdisciplinary projects and the application of software engineering to solve real-world problems.

• LinkedIn: linkedin.com/msgabriel
• Email: gabriels.msousa@gmail.com

Feel free to explore my repositories and projects. I'm currently building my portfolio with my professional experiences and knowledge. I am always open to new opportunities and collaborations. Let's connect and build something amazing together! 🚀