Davi Prata

This is currently a single-person independent research project led by Davi Prata, an AI/ML developer and data science practitioner based in Brazil. I hold an MBA in Data Science and Analytics from USP/ESALQ, where my capstone project received the highest grade and focused on pattern recognition using machine learning models applied to accelerometer signals. My relevant expertise is primarily technical: experimental design for model evaluation, Python-based data workflows, machine learning, reproducible logging, and development of software pipelines for controlled data collection. Over the past months, I have been building a custom evaluation framework to run standardized LLM benchmarks with traceable prompts, model configurations, outputs, and execution metadata. The scientific domain of this project is medical education and AI evaluation in a Brazilian context. I am not a physician and do not intend to make clinical claims. The project is designed as a benchmark study: it evaluates how LLMs perform on standardized medical education questions and analyzes their limitations, error patterns, and variability. The goal is to produce evidence that can support more responsible discussion of LLM capabilities in health-related domains, especially in countries where local benchmarks and public evidence are limited.

Personal Website

https://github.com/RockManRK/

LinkedIn

LinkedIn

Projects

No grants recorded.