Rafael Souza Pinto

Synthetic data and anonymization for financial and HR data, including TabDDPM, NER-based PII removal for Portuguese text, and stochastic differential equations for generating interest rate, FX and inflation scenarios. ML for recruitment, like hiring probability across pipeline stages and candidate matching. Also into privacy-preserving LLM pipelines with local models as extractors, and hybrid quantum-classical ML experiments. Background in low-latency infrastructure for Brazil's equities market.