I am a PhD Candidate in Computer Science (Artificial Intelligence) at ICMC-USP, working at the intersection of machine learning, relational learning, graph neural networks, and AI for healthcare.

My current research investigates how relational structure between participants can improve learning from small, heterogeneous clinical datasets, with particular interest in population similarity graphs, inductive graph inference, and rigorous evaluation of machine-learning models.

• Graph Neural Networks & Relational Learning — modeling participant relationships through population similarity graphs.
• Reliable Machine Learning for Small Data — leakage-controlled validation, robustness analysis, and careful model comparison.
• AI for Healthcare — developing methods that are useful under heterogeneous, incomplete, and data-scarce clinical settings.
• Model Understanding — studying feature importance, class-specific behavior, and the conditions under which relational models provide measurable value.

I hold a BSc in Biomedical Engineering from the Federal University of Uberlândia, including academic mobility at Polytech Marseille / Aix-Marseille Université. I previously conducted research in computational neuroscience and fMRI analysis at the University of Bristol and worked as a Junior Data Scientist at unico IDTech.

• Academic Website — source code for my academic website.
• GNN from Scratch — educational implementation exploring graph neural-network fundamentals.
• Undergraduate Research / TCC — code associated with my undergraduate research in neuroimaging and correlation analysis.

I am currently reorganizing this GitHub profile around reproducible research, educational graph-learning implementations, and reliable machine-learning workflows.

• Academic Website
• ORCID
• Google Scholar
• Lattes Curriculum
• LinkedIn

For research collaborations or academic inquiries: brunacampos.g@usp.br