Hiring · AI Engineer
Senior AI Engineer who ships LLM features to production
I am a software engineer first, which is why my AI work survives real users. I build LLM integrations, agents, and automation pipelines with the same care as any production system: typed code, logs, fallbacks, cost control, and a clear owner.
The gap I close
Prototype to production
Most AI projects stall after the demo. Getting them live needs data handling, error paths, evaluation, and integration with the existing stack.
AI without a business case
Models are cheap to call and expensive to maintain. I start from the process and the metric, then pick the simplest thing that moves it.
Teams that are new to AI
I write code the rest of the team can maintain and I explain the trade-offs, from prompt design to model choice.
What I build
LLM integrations
OpenAI, Gemini, and Perplexity APIs inside web products, with structured outputs and guardrails.
Agents and pipelines
LangChain agents and multi-step pipelines that read, decide, and act on business data.
Workflow automation
n8n and Make workflows that replace manual, repetitive work across CRMs, sheets, and messaging apps.
AI content systems
Automated content pipelines, including the agent that writes drafts for this blog from my daily work.
AI-assisted engineering
Daily use of Claude Code to ship faster without lowering the quality bar.
Full-stack foundation
React, Next.js, Node.js, and Rust, so the AI feature lands inside a real product.
Experience behind it
- AI Engineer at SINOON since 2025: automations and integrations with n8n, LangChain, Make, OpenAI, Gemini, and Perplexity.
- Front-end and product work for bMind.ai, an AI system for market and financial predictions.
- AI-assisted engineering with Claude Code at Peaksware (TrainingPeaks, MakeMusic).
- Instructor of a 16-hour hands-on AI workshop for business owners in Vitória, Brazil.
- Bachelor's degree in Computer Engineering and 6+ years in software engineering.
Related work
How I approach an AI feature
- 01
Process first
Map the task the AI will take over and the metric that proves it worked.
- 02
Smallest pipeline
Ship a narrow version with real data, logging, and a human in the loop.
- 03
Evaluate
Measure quality and cost on real cases before widening the scope.
- 04
Harden
Add fallbacks, monitoring, and documentation so the team can own it.
Stack
Frequently asked questions
What kind of AI engineering roles are you looking for?
Applied AI roles: building LLM features, agents, and automations inside products. Remote, full-time or contract.
Do you train models from scratch?
My focus is applied AI: integrating foundation models, building agents and retrieval pipelines, and shipping them as reliable product features.
Which tools do you use?
OpenAI, Gemini, Perplexity, LangChain, n8n, Make, and Claude Code, on top of TypeScript, Node.js, and Next.js.
Can you work in English?
Yes. I work in English and Portuguese and have delivered for companies in Brazil, Portugal, and the US.
Building an AI team?
Send me the role. I can walk you through how I would ship the first AI feature in your product.