Gabriel Gutierrez
Deep Learning,
Software Engineering
& Artificial Intelligence
Meet Gabriel
About Me
Machine Learning
I developed a passion for machine learning during my university studies, where I explored its vast potential to solve real-world problems. This interest led me to work as a junior researcher volunteer for two years, focusing on natural language processing projects. During this time, I conducted research on fake news detection using BERT, deepening my understanding of cutting-edge AI models.Currently, I am pursuing a master's degree at Unicamp, where my work focuses on the semantic segmentation of seismic facies. I had the opportunity to present my findings at the 85th annual EAGE conference and exhibition, where I published my research, contributing to advancements in the field of geoscience.
Software Engineering
I hold a degree in Software Engineering from one of Brazil's top universities, where I had the opportunity to contribute to impactful projects. During my studies, I worked on two projects focused on medical data collection for research that aimed to inform public policy decisions. These experiences honed my skills in developing practical solutions for real-world problems.Since joining Unicamp's master's program, I've applied my software engineering expertise to research, bringing a unique perspective to tackling complex challenges. I currently lead Minerva, an open-source machine learning training framework designed to meet the specific needs of researchers, developed in collaboration with Discovery Unicamp. As the software architect, core developer, and core maintainer of Minerva, I ensure the framework is robust, flexible, and aligned with the evolving demands of the research community.
My Skils
With a strong foundation in software engineering and machine learning, I bring a unique blend of technical expertise and research experience to my projects. My background spans developing data-driven applications, contributing to impactful research, and leading open-source initiatives. I specialize in computer vision, semantic segmentation, and building scalable machine learning frameworks like Minerva. My skills extend beyond coding; I architect solutions that bridge the gap between cutting-edge research and practical implementation, always with a focus on innovation and collaboration.Transformers
I have extensive experience working with transformer models, having started with BERT during my research on fake news detection. In my master's program, I expanded my focus to include vision transformers (ViT) and multi-scale vision transformers (MiT) for semantic segmentation. I worked with state-of-the-art models such as SegFormer, Segmenter, and SetR to tackle the challenging task of segmenting seismic facies. My research also explored the comparative performance of transformer-based models versus traditional CNN architectures, which culminated in a paper that highlighted the advantages and limitations of each approach in geoscience applications.Computer Vision
My experience in computer vision spans various domains, with a strong focus on image processing and semantic segmentation. In addition to working with cutting-edge transformer models for segmentation, I've gained expertise in self-supervised learning (SSL) techniques, which are crucial for enhancing model performance with limited labeled data. Through my work on seismic facies segmentation and other image-based tasks, I've developed a deep understanding of the intricacies of visual data, enabling me to tackle complex problems with innovative solutions.NLP
I have a strong background in Natural Language Processing, having worked with advanced models like BERT to tackle real-world problems such as fake news detection. My work in this area involved leveraging pre-trained transformer models to understand and classify text data, contributing to the development of robust systems for detecting misinformation. This experience gave me a solid foundation in NLP techniques, including tokenization, attention mechanisms, and fine-tuning models for domain-specific tasks, allowing me to efficiently handle complex text processing and analysis challenges.My Work
Semantic
Segmentation
In my work on semantic segmentation, I have focused on applying advanced machine learning techniques to the field of seismic facies analysis. My master's research at Unicamp explored the use of transformer models like ViT and MiT to improve the accuracy and efficiency of seismic facies segmentation, comparing their performance with traditional CNN-based approaches.This research was published and presented at the 85th annual EAGE conference and exhibition, where it garnered significant attention. Both the paper and my presentation received some of the highest ratings from the conference reviewers, recognizing the impact of this work on the geoscience community. My contributions not only advance the application of machine learning in geophysics but also provide valuable insights into the effectiveness of transformer models for complex geospatial tasks.
Natural Language
Processing
My work in NLP has focused on developing solutions to address pressing issues in information integrity, particularly in the detection of fake news. During my time as a junior researcher, I utilized BERT, a leading transformer model, to design a system capable of identifying misinformation with high accuracy. This research involved fine-tuning the model for the specific task of text classification, optimizing it for detecting patterns and nuances in language that signal deceptive or misleading content.This work laid a strong foundation in NLP, allowing me to contribute to projects requiring deep understanding of language processing and the application of cutting-edge transformer models in real-world scenarios.
Minerva
As the lead of Minerva, an open-source machine learning training framework developed at Discovery Unicamp, I've played a key role in its architecture, development, and maintenance. Minerva was designed to meet the specific needs of researchers, providing a flexible and scalable solution for training and fine-tuning machine learning models across a variety of tasks. My role as the software architect, core developer, and core maintainer has involved ensuring that the framework is robust, user-friendly, and adaptable to the fast-evolving demands of machine learning research.Minerva stands out by offering an accessible and powerful tool for both academic and industry researchers, bridging the gap between research and practical implementation. Through my work on this project, I've gained valuable experience in open-source software development, contributing to a tool that empowers the research community to advance machine learning in meaningful ways.
