**You will build the agent that UTUA will trust to make decisions about real revenue.** This is not prototyping. This is not a demo. This is production AI engineering—with evaluation, guardrails, provenance, and quality discipline—for an agent that analyzes monetization across ~50 countries and initiates actions with direct financial impact. UTUA is rebuilding its operations to be AI-first. The **Cortex & Sinapse** platform makes this possible. You will be the team’s AI agent engineering reference and the one who defines the AI-leverage maturity standard for everyone. **What you will build:** * Production UTUA Analyst Agent, with harness, interchangeable skills, evaluation, and episodic memory * Semantic design of Sinapse’s MCP tools—tight contracts that the model understands and uses correctly * Evaluation framework that measures quality regression for every prompt or skill change * HITL workflows and guardrails for write actions with financial impact **What we’re looking for:** * Experience building production LLM-based AI agents—not demos (tool use, harness, context management) * Evaluation discipline: evaluation, correctness criteria, regression detection * MCP tool design with tight contracts and awareness of god-tool * Active and mature use of generative AI in development—you are the team’s reference in this area **Contract model:** PJ **Compensation:** To be discussed
João Silva
Indeed · HR
João Silva
Indeed · HR
