Analytical AI professional with a multidisciplinary background in Computer Science and Law, specializing in RAG, LLM, and semantic search solutions. Fluent in English and Portuguese.
I'm an AI Developer and Data Analyst combining deep technical expertise with a unique legal background — giving me a rare ability to navigate complex, data-driven problems with precision and strategic thinking.
My work focuses on retrieval-augmented generation (RAG), LLM pipelines, and semantic search using Python, R, and SQL. I've worked across software QA, database development, and machine learning in fast-paced environments across Brazil and Canada.
I thrive at the intersection of cutting-edge AI tooling and real-world applicability — building systems that are not just intelligent, but practical and impactful.
Analyze and interpret data sets with preprocessing pipelines; build data-driven insights using AI/ML techniques; leverage AI models to extract information from datasets; create and present stakeholder recommendation reports.
Designed and executed front-end and back-end tests for a web application connected to REST APIs and SQL databases. Reported results and supported analysis within the Microsoft Azure environment using data visualization tools.
Managed high-volume usability and acceptance testing for 122+ mobile devices daily, coordinating with international carriers and presenting results in weekly global meetings. Led a regional team updating Latin American user manuals.
Managed usability and acceptance testing for smart devices in collaboration with major carriers (Vivo, TIM, Claro). Streamlined technical documentation, analyzing and refining up to 13 mobile user manuals per week during peak periods.
Provided technical support diagnosing and resolving hardware, software, and network issues — resolving up to 17 tickets per day. Assisted with system maintenance, user account management, and troubleshooting.
Designed a SQL database to store records for ~200 low-income families. Developed a login-enabled interface using HTML, CSS, and JavaScript to securely register and manage beneficiary information.
Provided legal representation support to low-income and marginalized populations unable to afford legal counsel, covering civil and criminal cases.
Developed a retrieval-augmented search algorithm converting recipe CSV data into FAISS vector indexes using Sentence-Transformers embeddings, with preprocessing and chunking in pandas for semantic search with dietary and recipe filters. Implemented metadata alignment and reproducible manifests for model tracking.
Applied AI classification algorithms — Logistic Regression and Neural Network — to predict heart failure using a clinical dataset with 12 features and 918 observations. Focused on model interpretability and performance benchmarking across both approaches.