Back home
Portrait of João Augusto

João Augusto

Senior Software Engineer · Brazil · Remote

Download résumé (PDF) Two-page PDF, ready to share with your team.

About

Senior Software Engineer with 6+ years of experience building distributed, event-driven systems in .NET for large companies in online dating and pharmacy retail.

My core stack is C#, .NET and Kafka, and I also work with RabbitMQ, Redis, SQL Server, Docker, Kubernetes, AWS, OpenTelemetry, Grafana, React, TypeScript and AI tooling (LLMs, MCP).

At Match Group, I work on the notification system behind email, SMS and push for 17+ brands, including Match, OkCupid, Meetic and Plenty of Fish, sending over a billion messages a month.

Often described by colleagues as a proactive team player with strong ownership, clear communication and a knack for solving complex problems.

Skills

Proficient

C#, .NET Core, TypeScript, SQL Server, Node.js, JavaScript, React, PostgreSQL, Docker, RabbitMQ, Redis, Entity Framework, Dapper, Git

Intermediate

Kafka, AWS (EC2, SQS, Lambda, S3, DynamoDB), Kubernetes, OpenTelemetry, Grafana, Prometheus, Loki, Mimir, CI/CD (Jenkins, GitHub Actions, Azure Pipelines), xUnit, Jest, Hangfire

AI & tooling

LLMs, MCP, Claude, Cursor, AI-assisted validation and spec-driven development

Practices

SOLID, Clean Code, Clean Architecture, Spec-Driven Development, Scrum, Kanban

Experience

Senior Software Engineer · Match Group

Dallas, TX, United States (Remote) · A global leader in online dating and social discovery, owning brands such as Tinder, Match, Meetic and OkCupid.
  • Reduced notification processing latency by 60% (from 2.5s to 1.0s avg) and eliminated database contention under peak load by migrating from a SQL Server queue to a Kafka-based event-driven architecture in C# and .NET.
  • Lowered Redis master CPU utilization by ~60% (from ~80% to ~30%) and improved notification processing latency by 75% (from ~2s to ~500ms) by distributing reads across replica nodes and caching recipient eligibility and template rendering.
  • Brought the mean time to detect critical incidents under 5 minutes by designing 64 alert rules across 5 backend services and a self-service Grafana dashboard with real-time delivery visibility across 15+ brands.
  • Cut observability ingestion costs by ~80% across 2 backend services by auditing and removing high-volume, low-signal log events.
  • Brought post-deployment regression detection under 1 hour across 6 backend services by building AI-powered validation skills that integrate the Grafana MCP, database health checks and pull request correlation.

Software Engineer · Pague Menos

Brazil (Hybrid) · One of the largest pharmacy chains in Brazil, with a strong national presence.
  • Cut monthly operational costs of warehouse labeling and distribution by 50% (from ~$200k to ~$100k) by modernizing a legacy labeling system with C#, .NET 8, React, TypeScript and SQL Server.
  • Reduced system downtime by 60% (from ~5h to ~2h) with Kubernetes health checks and stronger service resilience using .NET Core and distributed messaging patterns.
  • Halved incident resolution time (from ~2h to ~1h) by implementing OpenTelemetry-based observability.
  • Boosted system throughput and reliability by 40% by introducing asynchronous communication between services with RabbitMQ.

Software Developer · NDS

Brazil (Hybrid) · A technology center within a Brazilian Federal Institute, delivering IT services to public sector institutions.
  • Cut time spent on manual academic and administrative processes by 70% (from ~10 to ~3 hours) for 4,000+ users by building a full academic ERP system with C#, .NET Core, TypeScript, React and MySQL.
  • Shortened build times by 30% (from ~10 to ~7 min) by adopting an NX monorepo architecture.
  • Improved frontend data-loading time by 50% (from ~250ms to ~120ms) with React Query caching integrated with .NET Core.

Education

Bachelor's Degree in Computer Science
University of the Americas, Brazil
2026

Languages

English · Native or bilingual
Portuguese · Native
Spanish · Basic, fluent reading

Résumé updated November 2026