Where is Vinícius Raposo based?

Vinícius Raposo is based in Brazil and works with international clients. His collaboration window is Brazil time, UTC−3, which overlaps with the Americas and part of Europe.

About Me

I fix AI systems.
They should not be broken.

Q: What does Vinícius Raposo specialize in?

Vinícius Raposo specializes in production AI infrastructure: RAG reliability, AI agent workflows, LLM application development, data ingestion, evals, and AI system audits for founders and small teams.

Q: How does Vinícius work with clients?

Vinícius works async-first through clear fixed scopes. He explains the business problem, the technical failure mode, the implementation plan, and the tradeoffs in plain language before building or recommending fixes.

70% of RAG systems fail in production. I build the 30% that don't. AI infrastructure that actually works — not demos, not prototypes. Systems that ship value while you sleep.

Background

I'm Vinícius Raposo. Based in Brazil. I build AI infrastructure that survives contact with reality — the gap between "it works in the demo" and "it works on Tuesday at 3am when the edge case hits."

I came up through software engineering, then drifted into AI because that's where the interesting infrastructure problems are. These days I live in the messy middle: pipelines, retrieval systems, agentic workflows, and the deployment architecture that makes AI actually useful for businesses with real constraints and real users.

What I actually do

Clients come to me when their RAG system started hallucinating in production, their AI agent workflow became a maintenance nightmare, or their automation pipeline couldn't handle edge cases. I figure out what's wrong, redesign it properly, and ship a version that works.

I work mostly with small teams and founders. That's where AI has the most leverage — and where "good enough" is usually the standard when it shouldn't be. A well-built internal assistant replaces hours of daily work. A bad one just adds overhead.

How I work

Async-first, plain language, no jargon. I explain what I'm building and why. If a simple solution solves the problem, that's what you get — I don't upsell complexity.

Every engagement starts with understanding the business problem. The AI is the implementation, not the goal.

Core Skills

RAG Pipeline Design & Reliability
Agentic Workflow Architecture (Mastra, LangGraph)
LLM Application Development
Data Pipeline Engineering
AI System Evaluation & Audits
Content Automation Systems

Find me at

Short version.

The answers clients usually need before they decide if I am the right fit.

What do I specialize in?

Production AI infrastructure: RAG reliability, AI agent workflows, LLM application development, data ingestion, evals, and AI system audits for founders and small teams.

How do I work?

Async-first, fixed-scope, and plain language. I explain the business problem, the technical failure mode, the implementation plan, and the tradeoffs before building.

I fix AI systems.They should not be broken.