Building production-grade ML and Generative AI systems that deliver measurable business impact.

Machine Learning Engineer Specialist | Marlabs

Responsible for the architectural design, technical leadership, and implementation of Machine Learning and GenAI solutions in the healthcare domain, working directly with PracticeTek as technical reference for the initiative. The project focuses on Revenue Cycle Management (RCM), reducing claim rejections and denials through predictive modeling and intelligent automation, directly impacting financial efficiency and operational performance.

• Designed and implemented Machine Learning models for predicting claim rejections and denials, enabling proactive intervention and significantly improving claim acceptance rates and revenue recovery.
• Led the development of GenAI-driven solutions using LLMs to generate deterministic rules for error prevention, combining statistical learning with rule-based systems to increase explainability and reliability in production.
• Architected and deployed the end-to-end solution on AWS (SageMaker and Bedrock) and Snowflake, ensuring scalability, performance, and seamless integration with healthcare data pipelines, with full observability through Datadog.

Data Scientist | SOUTH SYSTEM

Responsible for architectural design, technical decision-making, and implementation of ML and GenAI solutions for the internal job portal and enterprise clients, leading initiatives that connected technical innovation to direct business outcomes.

• Architected an intelligent resume-to-job matching system on AWS with RAG, Pinecone hybrid search, metadata filtering, and Cross-Encoder reranking, contributing to roughly 50% revenue growth.
• Built a complete MLOps pipeline on SageMaker for oncology symptom-criticality prediction, with drift detection and automated retraining triggers.
• Developed an AI agent using Bedrock AgentCore with MCP integration via AgentCore Gateway, scaling to hundreds of concurrent users with Langfuse observability.
• Led a RAG-based GenAI solution for Grupo Santander on Azure, recognized as a Top 3 solution in an international innovation challenge.

Machine Learning Engineer | SantoDigital

Developed AI and MLOps solutions on Google Cloud Platform for internal products and enterprise clients.

• Built a clustering solution on Vertex AI with automated preprocessing and Markov Chain analysis integrated with BigQuery, surfaced through a Streamlit dashboard on Cloud Run.
• Led core architectural enhancements to an internal multi-agent product, expanding agent capabilities and scalability.
• Delivered a multi-agent system for an automotive financing client using MCP servers for context-aware customer service workflows.
• Fine-tuned models on Vertex AI with RLHF to align response tone and style to different business contexts.

Machine Learning Engineer Intern | SantoDigital

Built CI/CD retraining pipelines and GenAI chatbots with RAG and function-calling patterns on GCP.

• Replaced traditional NLP methods with multimodal text and image embeddings, improving accuracy and efficiency.
• Designed automated MLOps pipelines with CI/CD, streamlining retraining of business-critical models.
• Created a FastAPI and GenAI application for automatic product description generation, reducing operational costs and increasing sales volume.

Scientific Researcher | CNPq / SSP-PI Project

Developed Generative AI and NLP solutions to optimize management of Public Security Secretariat data, including intelligent conversational agents for automated 190 emergency service attendance.

MLOps & LLMOps Specialization | Duke University

Advanced specialization covering DevOps and DataOps for ML, pipelines, model deployment, MLflow, Hugging Face, cloud ML platforms, and LLM fine-tuning and deployment in production systems.

BSc Computer Science | Federal University of Piaui (UFPI)

Research activity in the CEREIA project on quality interfaces for health data science, NLP research with the Public Security Secretariat, and scholarship monitor for Digital Circuits.