A short page about the person
behind the notes.

I am a PhD student in Computer Science at UNICAMP, in Campinas, Brazil. This blog is an extension of my academic life, but behind the papers, code, and experiments there is someone genuinely curious about how complex systems learn, fail, and adapt.

My background is in machine learning and computer vision, with a few years of applied research at the intersection of theory, engineering, and the messy parts of real data. During the doctorate I have leaned more and more into semantic segmentation under weak supervision, trying to reduce the reliance on dense, fully annotated datasets while still producing robust, meaningful representations.

A lot of my taste was shaped by working with seismic data, specifically facies segmentation. Seismic teaches you, fairly quickly, that data is noisy, labels are expensive and imperfect, and evaluation is rarely as clean as benchmark numbers suggest. I expect to keep using seismic as a recurring testbed.

I am also the creator and technical lead of Minerva, an open research framework developed within my group to support reproducible ML experimentation and scalable workflows. It reflects my interest in clean abstractions, extensibility, and the practical realities of running large experimental pipelines.

I work in Portuguese and English, with Portuguese being my native language. I also study Japanese (JLPT N5, slowly working toward N4), partly as a long-term personal project, partly because it is nice to think in a different grammar once in a while.

Get in touch

I like talking about research, formally or otherwise. If something here resonates, if you are working on similar problems, or if you would just like to compare notes, please feel free to reach out.