Agentic AI Engineer

Overview Item.com is seeking an Agentic AI Engineer to build, deploy, and operate reliable AI agents that interact with business systems, data, APIs, and human operators.

Responsibilities

• Create agentic systems that safely complete multi-step work.
• Make agent behavior observable, measurable, recoverable, and cost-conscious.
• Turn real business workflows into reliable agent-assisted capabilities.
• Build agents using tools, APIs, retrieval, structured outputs, state, and human approval paths.
• Design execution flows with clear permissions, validation, retries, timeouts, escalation, and recovery behavior.

Minimum Requirements:

• Implement agent evaluation, tracing, observability, cost controls, and operational metrics.
• Develop secure tool interfaces and control access to data and downstream actions.
• Analyze agent failures and improve prompts, tools, policy, workflow design, and fallback paths.
• Partner with product and operations teams to define safe autonomy boundaries.
• Demonstrate ownership of production AI/LLM applications with agentic behavior.

Preferred Requirements:

• Strong Python and/or TypeScript engineering capability; experience with APIs, backend services, cloud, and data stores.
• Practical experience with tool calling, retrieval, state, orchestration, evaluation, and production observability.
• Ability to explain permissions, human review, hallucination containment, tool failure, and incident response.
• Framework familiarity is useful, but evidence of engineering and operational ownership matters more.