I hold a PhD in Electrical Engineering from UFMG, with research focused on mathematical modeling of time series. I've worked on mathematical modeling since 2007 and have been a professor and researcher at UFSJ since 2013.
My work spans system identification, data-driven modeling and control, and machine learning applied to physical systems.
Português · Projects · Research · Beyond work · LinkedIn
From the bench to Python. Data acquisition and experimental analysis with Arduino/NI-DAQ, plus data-driven modeling and control.
I created and wrote the original implementation. I proposed, supervised and tested later extensions implemented by students.
Docs · Examples · JOSS paper
Models from data. System identification with NARMAX models.
Started in a master's course I taught. I supervised the initial work, contributed code and scientific guidance, and coauthored the software paper.
Docs · Examples · JOSS paper
Financial data, via Telegram. A bot I developed for automated technical, fundamental and statistical analysis.
Private source code; I currently run it on my own computer. Public documentation includes commands and example outputs.
User guide · Documentation repo
Learning by working through examples. Course materials and notebooks on regression, classification and model validation in engineering.
A teaching repository, primarily in Portuguese.
Starting with hardware? There's also an introductory Arduino and NI-DAQ guide in Portuguese, which I coauthored.
- Learning from data: regression, classification, preprocessing, regularization and model validation.
- Understanding dynamics: nonlinear identification, NARX/NARMAX, hysteresis, control and multiobjective optimization.
- Knowing when a model is trustworthy: free-run simulation, finite-precision effects and interval methods.
I'm also interested in applying ML, data science and system identification to financial time series.
Python is central to my scientific software work. My programming background also includes C, C++, C#, Fortran, Pascal, MATLAB, Scilab and Octave, plus HTML and CSS for web work.
I've been a member of GCoM, the Control and Modeling research group at UFSJ, since its founding in 2007, initially as a student, and have led the group since January 2025. The group brings together students and faculty. I have managed research teams of undergraduate and graduate students.
My academic record includes 25 journal articles and 51 full conference papers. I have supervised dozens of undergraduate and graduate students.
I'm also a co-inventor of a granted Brazilian patent for automated human gait analysis (BR 10 2017 008238-5).
Marathon runner, Taekwondo practitioner and RPG completionist. I like long-term challenges and seeing things through — whether that's a finish line or the last side quest.
I also develop VibeLight-Tizen, a Moonlight Tizen fork for game streaming on Samsung TVs, with extensive changes to its audio, video and user interface.