Applied AI Engineer

Mizo builds AI agents that automate service desk operations for Managed Service Providers. The Applied AI Engineer will design, ship, and operate production agentic AI systems, taking agents from prototype to production and ensuring they behave reliably with real customer workloads.

Responsibilities:

• Build production agentic systems : design multi-step agent loops with tool use, planning, memory, and autonomous decision-making that run reliably against real customer workloads
• Own agent reliability end-to-end : instrument, monitor, and debug agents in production — chase down hallucinations, failed tool calls, runaway loops, and edge cases, then ship fixes
• Build evals and guardrails : create evaluation suites, regression tests, and safety checks so we can ship agent changes with confidence instead of vibes
• Optimize cost, latency, and quality : make data-informed tradeoffs across models, prompting strategies, retrieval, and orchestration to keep agents fast, cheap, and accurate at scale
• Integrate agents with real systems : connect agents to MSP tools, ticketing platforms, and internal APIs so they can take real actions, not just generate text
• Close the loop with production data : turn real agent runs and user feedback into better prompts, better tools, and a better roadmap
• Collaborate async-first : communicate clearly in writing, share context proactively, and make decisions with minimal coordination overhead