Fernando – Node.js, PostgreSQL, Typescript

A security log platform for managed service providers (MSPs and MSSPs) that need to detect threats across many customer environments without running their own security setup. It collects and processes logs from customers' firewalls, endpoint agents, and cloud accounts in near real time, with high availability and low latency. The product is also sold through a third-party IT marketplace, where providers can onboard their customers themselves and receive a single bill for all their IT products. The platform handles a baseline of about one million events per second, or hundreds of millions of events a day.

• Found the most expensive Kafka consumers in the Confluent bill and fixed the cause in a slow Node.js consumer by batching its Redis calls and stopping it from blocking the event loop, cutting that workload from 120 to 50 pods. This was part of a team effort that cut infrastructure costs by over $200K a year.
• Designed, led, and launched the award-winning Pax8 marketplace integration end-to-end, from scoping with product and the partner team to production support; self-service onboarding and consolidated billing drove 5% growth in the customer base.
• Built the company-wide Integration Framework for data ingestion, cutting integration development time by 60% and solving eventual-consistency problems with CQRS, Kafka, and S3.
• Built an event-driven HubSpot CRM sync that kept account managers' revenue view current when customers changed their own contracts, improving MRR forecasting accuracy by 20%.
• Rewrote the payment module from backend to frontend on Stripe's newer APIs, adding bank debit (ACH) and other payment methods and reworking the payment data model.
• Played a key part in migrating Kafka from AWS MSK to Confluent, and fully owned the cold-storage layer of the Spark-based data platform.
• Moved a mission-critical service off a legacy system, protecting key customer flows from its downtime, and built Prometheus alerting and Grafana dashboards that cut incident response times.

A large education platform that serves learning materials to schools, universities, and their students. Its core services handle authentication and authorization for all public and internal endpoints, as well as search across the course material catalog, organizations, classes, and enrollments.

• Optimized the Elasticsearch cluster behind catalog and enrollment search by replacing frequent real-time document updates with scheduled batch updates (Painless scripts), cutting query latency by 25%.
• Owned the authentication and authorization services protecting all of the company's public and internal endpoints.
• Rewrote the API test codebase, raising test coverage by 30% and adding unit, integration, end-to-end, and load testing (JUnit, Cucumber, k6, Jest, Playwright).
• Reviewed and guided contributions to core features in Go, Node.js, Spring Boot, and Angular.
• Built Prometheus and Grafana dashboards for real-time monitoring of the core services.

A payments and financial reporting system for commercial airlines that buy aircraft and services from a large aerospace manufacturer. It processes payments and generates financial reports for the enterprise's finance teams and airline customers, with a focus on smooth, reliable payments.

• Built a financial report generator API in Go that automated financial reporting for the enterprise's finance teams.
• Owned the payment flow across the frontend (React) and backend for commercial airline customers.
• Improved the payment microservices with Node.js streams, PostgreSQL transactions, and GraphQL to make payment processing more reliable.

A major Brazilian news and content portal with editorial, advertising, and subscription products for a large national readership. It runs on a mix of legacy systems and microservices that serve news content and paid features.

• Improved the news platform's Lighthouse performance and SEO scores by 30%.
• Led a small team that shipped a new payment feature in ASP.NET Core under a two-week deadline.
• Led the move of legacy systems and microservices to Docker containers.