Senior ML-focused systems analyst defining MLOps foundations and owning the lifecycle of NLP models in a large-scale justice environment.

• Reduced processing time by 25% by refactoring and productionizing legacy ML services used in judicial workflows.
• Enabled model adoption across more than 500 internal users by building Django APIs that embedded ML outputs into analyst workflows.
• Cut model refresh cycles from weeks to days by automating ETL, retraining, and release workflows with Airflow and CI/CD.
• Reduced manual classification effort by 85% by delivering an NLP assistant tailored to analyst review workflows.
• De-risked RAG adoption for legal retrieval by validating embeddings, vector search, and AWS integration paths for LLM pipelines.

Machine learning engineer delivering legal-tech R and D and production ML systems for Brazil's higher courts, connecting legal stakeholders, data workflows, APIs, MLOps, and technical delivery.

• Improved rollout readiness across judiciary programs by leading legal-tech discovery, KPI definition, and delivery planning with institutional stakeholders.
• Accelerated deployment of legal AI systems by shipping supervised, unsupervised, and semi-supervised models behind production APIs with standardized MLflow and DVC governance.
• Surfaced more than 30 precedent and semantic categories by designing topic-modeling and semantic-similarity pipelines for long-form judicial text.
• Reduced drift and annotation waste by implementing active-learning loops driven by user feedback and evolving legal language.
• Improved auditability and incident response by delivering end-to-end data, observability, and lineage workflows for production legal AI systems.

First industry experience as a full-stack developer in a consulting environment building and maintaining an internal Node.js and Vue.js intranet backed by MongoDB.

• Accelerated internal feature delivery by contributing across frontend and backend in a Git-based consulting workflow.
• Improved usability across devices by building responsive internal interfaces with Vue.js and Angular.
• Enabled authentication, CRUD flows, and business rules by implementing REST services in Node.js and Express.
• Stabilized application data flows and common-path performance by modeling MongoDB collections, indexes, and aggregation queries.

Centro de Ensino Médio Integrado do Gama (CEMI)

Biometric Payment System for Public Transport

Feb 2016 - Oct 2017 · Brasília, Brazil

Built during technical high school at CEMI, this Arduino and IoT prototype explored biometric bus-fare payment as a way to replace physical tickets and cards with fingerprint-based authentication.

• Integrated biometric authentication with a fingerprint scanner.
• Processed fare validation on Arduino hardware with embedded C.
• Explored IoT connectivity, RFID, and secure handling of biometric data.