Senior AI Engineer Agentic AI · RAG · Architecture
AI agents in production
Hi, I'm FILIPI.
Senior AI
Engineer
I put generative AI into production: agents with LangGraph, RAG, guardrails, evals and observability, wired into the company's systems. Testable, governed and live.
See projectsRecife, Brazil · Remote
New AI,
old discipline.
- 10+Years writing
code - 9Certifications in
AI, cloud & security - 9kLinkedIn
followers
Building software / 2016–2026
Work that runs itself.
No project with that name.
Built withThe everyday toolkit.
- Python
- LangGraph
- LangChain
- FastAPI
- Claude
- Postgres
- ClickHouse
- Docker
- Kubernetes
- Spring Boot
- React
- TypeScript
From sketch. To shipped.
FinOps for the cloud, end to end: a read-only Azure connector, per-organization isolation with row-level security, FOCUS-aligned cost data and explainable numbers, with the evidence behind every value.
Drag to see the processDrag / arrow keys
More than code.
Systems that work.
I'm Filipi, a Senior AI Engineer. I've been writing software since 2016, and today I design and evolve generative AI agents in large-scale enterprise environments. I bring software engineering, architecture and product thinking together so an LLM becomes a system: testable, observable, secure and governed in production.
Education & focus
Education
- Software Engineering2025 – 2028Bachelor's degree at UniFatecie, in progress.
- Systems Development2025 – 2026Technical degree from Globaltec, completed.
- Agentic AI in production2023 – nowPython · LangGraph · LangChain · Langfuse · FastAPI
Specialties
From prototype to production
- 01Problem
I map the flow, the data and where things can go wrong.
- 02Architecture
State, tools, memory and what never runs without a human.
- 03Prototype
An agent working with real data and structured outputs.
- 04Reliability
Guardrails, retries, timeouts, fallbacks and evals.
- 05Production
Tracing in Langfuse and data-driven iteration.
Autonomy isn't declared. It's earned with data.
AI that runs
in production.
Always building. Always live.
Experience & milestones
-
Technical reference in generative AI for Vivo: an agent platform for technical support, context routing and disambiguation, modular architecture, and observability with tracing and evals in Langfuse.
-
Enterprise applications with LLMs (OpenAI and Claude), RAG pipelines with semantic search, and APIs in Python/FastAPI and Java/Spring Boot integrated with microservices.
-
End-to-end generative AI solutions: RAG with vector databases, plus agents and assistants connected to APIs, documents and databases.
-
A design system in React, Next.js and TypeScript: 40% faster sprint delivery and a 15% improvement in Core Web Vitals.
-
Web applications, back-end services and REST APIs with Java, JavaScript and React, deployed and maintained in the cloud.
Written in production.
- 01Feb 19, 2026 Building Enterprise Payment Systems: Microservices, Saga Pattern, and Observability at Scale The reference project I spent three months building: a payment system with microservices, the Saga pattern and observability. Read on LinkedIn
- 02Jan 21, 2026 Engenharia de Performance: Como Otimizei um Roguelite Cross-Platform para 60 FPS Estáveis no Navegador From the first lag spike caused by the garbage collector to a steady 60 FPS, in a game that runs entirely in the browser. Read on LinkedIn
Let's talk
agents.
Agent architecture, RAG in production, or a good conversation about generative AI. Reach out here.
Start a conversation- Email[email protected]
- WhatsApp+55 81 99586-6387
- LinkedInin/filipimoraes
- GitHubFilipiWanderley
Based inRecife, Brazil · Remote
Built
live.