I combine generative AI, NLP and machine learning with the full data process behind data science: from structuring and data quality to modeling, automation and real-world applications. Experience in automated auditing, contract reading, clause extraction, compliance and technical information retrieval with LLMs and multi-agent systems.
Anti-hallucination with curated answers, hybrid search and model-level resilience. FNDE.
PySpark and Databricks to find inconsistencies and speed up auditing.
Contract reading, clause extraction and payment flow with AI.
Natural-language questions over a relational base, with semantic context.
Models fine-tuned in Portuguese for named-entity recognition in the legal domain (LeNER-Br). F1 of 83.76%, with an app published on Hugging Face.
Convolutional neural networks to classify chest X-rays supporting COVID-19 diagnosis. 98.4% test accuracy.