Software engineer at allu., Belo Horizonte
I build customer-facing software used by a million people a month.
Software engineer on the growth team at allu., working on the customer-facing applications used by 1M+ people every month. Before that: payment infrastructure at allu., MCP tooling at dti Digital. Studying Computer Science at PUC Minas.
Animated diagram: an agent reads a diff, plans scenarios, runs tests and passes. payments · pix checkout
- 1M+
- monthly users of the allu. apps I work on
- R$ 591K
- recovered in one month by a retry worker I built
- 98%
- fewer context tokens in MCP tool discovery
Software engineer with production experience across fintech, consulting, e-commerce and enterprise chatbot platforms. I like owning problems end to end: from the data model and the queue to the API and the UI. My focus is AI growth: agent tooling, LLM features and customer-facing product at scale. Recent work also included a payment infrastructure project.
- now
- Software Engineering Intern at allu., growth team
- education
- B.Sc. Computer Science, PUC Minas (2025–2028)
- focus
- AI growth, agent tooling, growth engineering
- languages
- Portuguese (native), English (C1, Cambridge)
- location
- Belo Horizonte, MG, Brazil
Education
- 2025 — 2028
- B.Sc. Computer Science, PUC MinasExpected graduation 2028
Awards
- Aug 2025
- Best Software Development Project, 1st place
- Jun 2025
- 3rd Unifecaf Hackathon, 3rd place
- Now on the growth team, building and maintaining allu.'s customer-facing applications, used by 1M+ people a month.
- Previously worked on payment infrastructure projects on a platform processing millions of reais monthly for a 60K+ customer base.
- Built a failed-charge retry system on Supabase Edge Functions as a serverless worker, batch-processing a PostgreSQL-backed queue. Recovered R$591,791.91 in a single month with no added infrastructure cost.
- Designed a proprietary white-label payment gateway with multi-tenant isolation and real-time webhook event delivery, reusable across client-facing products.
- Cut context token consumption in MCP server tool discovery by 98% by building MCP Optimizer, lowering API cost and chatbot latency.
- Designed and migrated FastAPI chatbot infrastructure from legacy routing to Model Context Protocol over JSON-RPC.
- Shipped production features on FastAPI and MCP used daily by 1,000+ people, reducing manual task time in internal operations.
Magic Mango
AI Engineering Intern
Dec 2025 — May 2026, São Paulo
- Automated lead generation and sales prospecting with n8n and Python/Go scripts, reducing manual outreach time by 90%.
- Built competitive-analysis tooling for prospects, turning market data into inputs for startup strategy.
- Built Java Spring Boot microservices for real client projects with a focus on maintainable architecture.
- Translated business requirements into technical specifications with multidisciplinary teams.
Click a row to open details
1,283 contributions in the last year
Hover a day for details
location unavailable (no geo headers here)
● live stats unavailable right now · amber pings are an illustrative animation · green is you
- languages
- Python, TypeScript, JavaScript, Go, Java, SQL
- backend
- FastAPI, Node.js, Spring Boot, GraphQL, REST, JSON-RPC
- frontend
- React, Next.js, React Native, Tailwind CSS
- data
- PostgreSQL, Redis, Supabase
- infra
- AWS (S3, CloudFront, CLI), Docker, GitHub Actions, CI/CD
- ai
- Model Context Protocol, Google ADK, Gemini, n8n
If you want to talk AI growth, agent tooling, growth engineering, or have something worth building together, email me.