name: Jose Santiago Rueda Antonio
role: AI Agent Engineer
location: Remote · Working with US-based teams 🌎
experience: 8 years shipping production systems
education:
- MSc. Artificial Intelligence — Tec de Monterrey (ITESM)
- MSc. Cybersecurity — Universidad Tecnológica de México
focus:
- Multi-Agent Systems & Agent Orchestration (LangGraph)
- RAG Pipelines & Retrieval Optimization
- MCP (Model Context Protocol) Tool-Calling
- Production Observability, Evals & Cost Control
status: Building production agent systems · Open for new contracts starting June 2026I architect systems where AI agents think, decide, and act autonomously. Multi-agent coordinators. RAG pipelines. LLM-driven UIs. MCP tool schemas. I've gone far beyond prompt engineering — I design the orchestration layer that makes AI production-ready.
"Anyone can call an API. I build the system that decides which API to call, when, and why."
The 3 things that separate a demo from a production agent — and what I bring on day one:
- Evals: LLM-as-judge + golden datasets + regression suite in CI. No agent goes live without a baseline I can compare against.
- Observability: full tracing with LangSmith / Langfuse — every tool call, every retry, every token cost is queryable.
- Cost control: semantic caching, model routing (cheap model first → escalate), structured outputs to cut retries. Typical savings: 40–60% of token spend.
| 🧠 | Designed multi-agent coordinator–specialist system with MCP tool-calling — routing user requests autonomously through specialized agents in production |
| 💹 | Built a financial AI agent (GPT-4 + LangChain) that analyzes dollar volatility and surfaces real-time transaction recommendations for a US fintech |
| 🧩 | Shipped LLM-driven UI orchestration — model selects and renders React/Next.js components as response output, replacing static text replies |
| 📡 | Built IoT device simulation framework (Python + FastAPI + MQTT) eliminating hardware dependency from QA cycles |
| 🔬 | MSc. Artificial Intelligence at Tec de Monterrey — graduating June 2026 |
If you're working on agentic AI, production LLM systems, or AI-powered products — let's talk.