AI Systems Engineer focused on hybrid architectures (LLM + deterministic logic) to eliminate hallucinations and ensure auditability in production.
Builds 100% local pipelines with R$ 0.00 infrastructure cost, mandatory Human-in-the-Loop and LGPD-by-design.
Method: Experiment first. Measure. Understand. Only then scale.
Principle: LLM is the last resource, not the first. If I cannot explain why a decision was made, I don’t have a system — I have a black box.
Experiment first. Measure. Understand. Only then scale.
RS4 Lab is the experimental lab of Rs4Machine — a space for controlled experimentation in intelligent systems, AI agents, and production-oriented engineering.
- 🧭 A human remains the final decision-maker on every critical call.
- 🛑 Every autonomous system needs a clear, fast way to be interrupted.
- 📊 Decisions driven by metrics, not intuition.
| # | Project | Description | Stack | Link |
|---|---|---|---|---|
| 01 | SofiaVoice v2.0 | Real-time voice assistant with local orchestration, speech recognition and natural conversation flow | Python · FastAPI · Ollama · Groq · Edge-TTS | 🚀 Repo |
| 02 | FraudEye | Fraud detection system for financial and invoice workflows using deterministic rules and model-assisted analysis | FastAPI · Python · Next.js · CV | 🚀 Repo |
| 03 | Translatia | Technical translation pipeline for documentation and structured content with language-aware processing | FastAPI · Python · Next.js | 🚀 Repo |
| 04 | FinAI | Financial assistant for analysis, planning and operational monitoring with local-first reasoning | Python · FastAPI · Next.js | 🚀 Repo |
| 05 | RS4Commerce Pipeline | Commercial opportunity triage pipeline focused on deterministic local workflows, auditability and low infrastructure cost | Python · Docker · Local AI | 🚀 Repo |
| 06 | RS4-cortex-flow | Local orchestration experiments with multi-agent coordination and memory, including LangGraph and persistent vector memory in active development | Python · LangGraph · ChromaDB · Ollama | 🚀 Repo |
Open to technical collaborations, AI systems consulting, and research-oriented product work.
📩 python.dev.raphael@gmail.com
🔗 LinkedIn
🌐 Portfolio
⭐ If any of the projects helped you, consider giving it a star.
Raphael Mendes · Rs4Machine · September 2026