Adriano Neves

I designed and implemented C programs for Monte Carlo simulations of spatial epidemic models on regular, irregular, and complex networks (Barabási–Albert, Watts–Strogatz, Erdős–Rényi, random geometric, and random spherical), identifying phase transitions and determining critical exponents in percolation theory. I optimized the computational performance of my simulations through parallelization with OpenMP and CUDA, reducing the runtime of large-scale simulations. I managed databases containing tens of terabytes by implementing the HDF5 format for efficient compression and access, drastically reducing the stored data volume and enabling scalable analyses.

I built a comprehensive data analysis system for influenza outbreaks in municipalities in Paraná using public data from OpenDataSUS, numerically solving systems of nonlinear differential equations with Python, and generating validated forecasts with an error margin of less than 5%. As part of a university extension program, I have been teaching physics and mathematics to talented/gifted students and mentoring them on undergraduate research projects since July 2024, which have earned awards at science fairs across the country.

Ph.D. Fellow in Physics at CNPq - UEM (2023-08 – Present)

I designed and implemented C programs for Monte Carlo simulations of spatial epidemic models on regular, irregular, and complex networks (Barabási–Albert, Watts–Strogatz, Erdős–Rényi, random geometric, and random spherical), identifying phase transitions and determining critical exponents in percolation theory. I optimized the computational performance of my simulations through parallelization with OpenMP and CUDA, reducing the runtime of large-scale simulations. I managed databases containing tens of terabytes by implementing the HDF5 format for efficient compression and access, drastically reducing the stored data volume and enabling scalable analyses.

I built a comprehensive data analysis system for influenza outbreaks in municipalities in Paraná using public data from OpenDataSUS, numerically solving systems of nonlinear differential equations with Python, and generating validated forecasts with an error margin of less than 5%. As part of a university extension program, I have been teaching physics and mathematics to talented/gifted students and mentoring them on undergraduate research projects since July 2024, which have earned awards at science fairs across the country.

Master's Degree student in Physics at CNPq - UEM (2021-08 – 2023-07)

I developed and optimized algorithms in C to simulate the spatial spread of diseases based on the SIR mathematical model and percolation theory, evaluating connectivity through short-range contacts in urban and agricultural settings. I applied high-performance computing techniques (optimization of data structures, loops, and memory access) to accelerate computationally intensive analyses, reducing processing time from months to days. I performed statistical analyses of the simulation results, determining critical thresholds for propagation and identifying phase transitions in the system.

I worked on predictive modeling to develop algorithms that use numerical methods to solve SIR-family mathematical models in Python, to analyze the spread of COVID-19 on a national and global scale, extracting epidemiological indicators and implementing a short-term forecasting system with a margin of error of less than 5%.

Institutional Scientific Initiation Scholarship at CNPq - UEM (2019-09 – 2020-10)

I developed a Python algorithm using machine learning (logistic regression) to classify artistic styles of paintings based on the Shannon entropy of pixel intensity patterns, combining image processing and statistical modeling.

Ph.D. in Physics – UEM (2023)

Master's Degree in Physics – UEM (2021 – 2023)

Bachelor's Degree in Physics – UEM (2017 – 2021)