At CCSC Research Lab, UFMG, I build and train deep learning models and investigate LLMs and NLP for semantic understanding and anomaly detection in data streams. I designed Python-based time-series embedding methods that reduced input data volume by up to 99% and training time by up to 98.9%.
I’ve operationalized data, training, and inference pipelines, and my research has resulted in three conference publications at IEEE ISCC, SBRC, and IEEE LATINCOM. At Pi Júnior, UFMG, I developed automated Python data pipelines; at iJúnior, UFMG, I worked as a Product Owner, prioritizing backlogs and facilitating Scrum teams.

