Generative AI Engineer · Brazil · Remote
I build LLM applications, RAG pipelines, AI agents, and backend systems that make model behavior measurable, observable, and ready for real users.
01 / Selected work
Open-source projects that show how I approach reliability, retrieval quality, local inference, and developer workflows.
A Python toolkit for detecting retrieval regressions before a knowledge-base or pipeline change reaches deployment. It compares versioned runs, reports query-level evidence, and applies a fail-closed policy when evaluation is incomplete or regresses.
View repositoryA privacy-focused terminal assistant for developers. Local inference through Qwen and llama.cpp, streaming output, persistent sessions, command validation, usage telemetry, and a tested Rich TUI.
View repositoryA local-first TUI for collecting job descriptions and generating evidence-based, AI-tailored PDF resumes. Profile data, job descriptions, LaTeX, and generated files stay local; applications are never submitted automatically.
View repository02 / Experience
Three years building AI products, backend systems, integrations, and delivery practices across Brazil-based teams.
Designing a multi-repository AI platform for aircraft maintenance and technical documentation. Building PDF/SGML ingestion, RAG, pgvector retrieval, LangGraph/DeepAgents workflows, AWS processing, APIs, streaming, evaluation, observability, and security practices.
Built conversational AI agents and enterprise CRM/ERP integrations with TypeScript, Node.js, LangChain, RAG, and MCP. Reduced agent response time by 40%, achieved 95%+ generated-response reliability, and mentored more than 10 developers.
Developed web applications and administrative dashboards, introduced end-to-end testing with Cypress, improved page-load performance by approximately 30%, and worked in Agile delivery cycles.
03 / Capabilities
The useful part is not the tool list. It is connecting the pieces into systems that can be tested, operated, and handed off.
LangChain, LangGraph, DeepAgents, MCP, RAG, agents, embeddings, reranking, structured output, prompt engineering, guardrails, and evaluation.
Python, FastAPI, TypeScript, Node.js, NestJS, REST APIs, WebSockets, PostgreSQL, pgvector, vector databases, and ingestion pipelines.
AWS, Bedrock, S3, SQS, Glue, DynamoDB, ECS, Docker, Kubernetes, KEDA, CI/CD, Langfuse, tracing, and token/cost accounting.
Testing, security hardening, Agile/Scrum, technical mentorship, enterprise integrations, POCs, documentation, and production reliability.
04 / Working together
I am based in São José dos Campos, Brazil, and work remotely. I am open to remote full-time, contract, or freelance roles and can operate through a Brazilian ME CNPJ for contractor engagements.
05 / Start a conversation