Gabriel Capiteli Bertocco

Gabriel Bertocco is a postdoctoral researcher in Large Language Models (LLMs) and Computer Vision at the Artificial Intelligence Laboratory, Recod.ai, Institute of Computing (IC), University of Campinas (Unicamp). From 2015 to 2018, he was an undergraduate research scholar in a project sponsored by Motorola, focusing on facial biometric analysis based on Deep Learning for age estimation in mobile devices. For this work, in 2017, he received the Inova Innovation Award for the best undergraduate research project in progress at Unicamp. He worked as a research scientist at SciPet Solutions in Technological Innovations during the first semester of 2019 and graduated with distinction in Computer Engineering later that same year. In 2024, he completed his Ph.D. in Computer Science with a focus on Artificial Intelligence and Computer Vision under the supervision of Prof. Dr. Anderson Rocha and Dr. Fernanda Andaló, funded by FAPESP. His doctoral research addressed the challenges of Unsupervised Re-Identification of People and Objects in multi-camera scenarios without labeled data. The focus was to develop self-supervised learning algorithms capable of handling biases, semantic variations, and different data modalities.During his Ph.D., he served as a Professional Research Assistant ( (2022-2023), at the University of Colorado Colorado Springs (UCCS, USA) and was a member of the Biometric Recognition and Identification at Altitude and Range (BRIAR) program (https://www.iarpa.gov/research-programs/briar), a U.S. government-funded project dedicated to counterterrorism, critical infrastructure protection, transportation facility security, military force protection, and border security. He published his proposed solutions in two first-author papers in the IEEE Transactions on Information Forensics and Security (T-IFS), with an impact factor of 8.0, one in the IEEE Transactions on Biometrics, Behavior, and Identity Science (T-BIOM), with an impact factor of 5.0, in addition to other publications and presentations in top-tier conferences such as the IEEE Workshop on Information Forensics and Security (WIFS), IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), IEEE International Joint Conference on Biometrics (IJCB), and IEEE Access. His Ph.D. defense committee included Prof. Dr. Patrick Flynn (University of Notre Dame, USA), Prof. Dr. Sébastien Marcel (IDIAP, Switzerland), Prof. Dr. Vitomir ?truc (University of Ljubljana, Slovenia), and Prof. Dr. Esther Luna Colombini (IC/Unicamp). In 2025, the Brazilian Computer Society (SBC) awarded his dissertation the Best Ph.D. Thesis in Computer Science in Brazil (2024). It was later also recognized as the Best Ph.D. Thesis in Computer Vision at SIBGRAPI 2025, one of the largest international conferences in Latin America on Computer Graphics and Computer Vision.He has also worked and continues to work as an AI consultant for national and international companies and government institutions, including the Dubai Police (United Arab Emirates), the Federal Legislative Police of Brazil, and several Public Prosecutors Offices. Gabriel Bertocco has over nine years of experience in machine learning, deep learning, and computer vision, with a strong record of publications, presentations, consulting projects, awards, interviews, podcast appearances, and a patent. He also serves as a reviewer for journals such as IEEE T-IFS, IEEE T-BIOM, and IEEE T-PAMI, as well as for major conferences. Finally, he collaborates on the Hórus Project, funded by FAPESP. (Source: Lattes Curriculum)