Roger Gelonezi

- Joining a new AI-first product team as Senior .NET Engineer, mapping the codebase end to end and surfacing architecture and process gaps directly to leadership.

- Partnering with the product team to reshape the roadmap, splitting features into parallelizable tasks and enabling small, frequent Continuous Delivery gated by code reviews, automated tests, and efficient AI token usage, allowing the team to ship new features faster.

- Driving engineering standards through RFCs and ADRs that shape team workflows and ecosystem-wide rules.

- Applying AI-assisted development practices to raise code quality and consistency across the team.

- Took ownership of a production Machine Learning product and raised classification accuracy from ~40% to 80%+, scaling a custom-trained, multi-layer classifier from 12 to 140+ labels for Brazilian legal documents.

- Automated case distribution within the public prosecutor's office, replacing manual triage with classification powered by custom-trained Machine Learning models.

- Redesigned a C# API connecting the Machine Learning models to a multi-tenant architecture, working toward a single high-confidence model deployable across all tenants.

- Simplified the workflow for training and testing custom models and authored full documentation, enabling any engineer to ramp up on the Machine Learning pipeline and its underlying concepts quickly.

- Deepened hands-on Data Science skills with MLflow and Qdrant to support reproducible, reliable delivery of custom-trained models.

- Architected a cloud native microservices ecosystem of 20+ C#/.NET APIs — built on Vertical Slice and Hexagonal Architecture with a shared SDK and event-driven communication via MassTransit and the Saga Pattern — processing 200,000+ legal proceedings per day, packaged into a single deployable product with full CI/CD.

- Built an AI-driven analysis routine integrated with Grafana that continuously tuned Kubernetes auto-scaling, cutting AWS resource reservations by 40% while eliminating API throttling under load.

- Integrated nearly 20 Brazilian judicial and prosecutor systems — including Tribunal do Trabalho, Tribunal Militar, SAJ, Eproc, Projudi, STJ, and DJE — enabling real-time synchronization between prosecutors and tribunals, with documents available via AWS S3.

- Established automated test gates (Robot Framework) that caught regressions at the pipeline level before release, replacing days of manual testing and protecting every deployment.

- Delivered end-to-end observability dashboards that surfaced integration issues before they reached production.

- Drove the four-year, progressive replacement of the legacy system, personally owning each client cutover with restricted-protocol overnight migrations that ensured zero-incident transitions until full legacy retirement.

- Mentored 5 engineers from intern to senior level — including an intern now working with AI in France — and led company onboarding to get new hires productive in a complex domain.

- Drove the migration analysis to modernize a legacy Visual Basic 6 monolith into a C# backend API with a React front end — mapping each module's risks and defining the execution plan, starting with the Orders module before its implementation was passed to a dedicated team.

- Built a React Native self-checkout app integrated with PagSeguro for in-store kiosks in bookstores — a success that PagSeguro featured as a benchmark case study.

• - Integrated an e-commerce platform with multiple marketplaces and built a reporting suite to support management decision-making.
• - Designed an end-to-end workflow controlling the full order lifecycle, from sale through delivery.

• Led technical management of a large-scale municipal system for the Bauru city government.

• Built software meeting ISO 10012 legal specifications for oil-extraction measurement with MODEC's metering team.

• - Modernized a legacy Visual Basic system into a C# .NET application.
• - In a small-business setting, wore multiple hats beyond engineering — handling client support and helping win new clients — building end-to-end product and customer experience early in my career.