π Based in Campinas, SP, Brazil
π PhD Student in Computer Science @ Unicamp
π§ Software Engineer | π€ Researcher in Machine Learning & Computer Vision
π Fluent in English and Portuguese | πΎ JLPT N5 Certified in Japanese
I am a researcher and software engineer with a strong interest in machine learning, computer vision, and self-supervised learning. My background combines solid software engineering foundations with applied research experience, allowing me to bridge the gap between theoretical advances and practical implementations.
π¨βπ I hold a Bachelorβs degree in Software Engineering and a Masterβs degree in Computer Science, where I specialized in deep learning and Transformer-based architectures.
π Currently, I am pursuing a PhD in Computer Science at the University of Campinas (Unicamp). My research focuses on:
- Efficient self-supervised learning (SSL) methods for vision and multimodal models
- Robustness and adaptability of foundation models
- Scalability and distributed training for large-scale models
π οΈ I am also one of the core maintainers and software architect of the open-source project Minerva, which provides scalable and flexible tools for deep learning research. My work on Minerva bridges research and engineering by enabling reproducible experiments, large-scale training, and model debugging.
πΎ In addition to my technical work, I also hold the Japanese Language Proficiency Test (JLPT) N5 certificate, reflecting my interest in languages and cross-cultural exchange.
- Self-Supervised Learning for vision and multimodal models
- Foundation models: robustness, adaptability, and efficiency
- Large-scale distributed training and optimization strategies
- Transformer architectures for computer vision
- Applied AI research at the intersection of software engineering and machine learning
I am open to collaborations and discussions on topics related to machine learning, self-supervised learning, and applied AI research.
π« Feel free to reach out via email at gabriel.bgs00@gmail.com or connect with me on LinkedIn.