Senior Agentic AI Specalist

Midwest Family Mutual Insurance Company•Urbandale, IA

1d•Remote

About The Position

The MFM IT team has a long history of delivering and maintaining core business systems in a mid-sized Property & Casualty insurance environment. As we expand our development team and adopt new AI development tools, we are making a deliberate investment in building this capability correctly from the start. As we begin to build agentic AI capabilities within our enterprise system, we are looking for an experienced engineer with strong knowledge of agentic AI systems to help guide this effort. The Senior Agentic AI Specialist will work closely with MFM's Principal Architect, who leads overall platform architecture, and partner with senior developers to design, implement, and improve agent-based solutions. This role will bring specialized expertise in areas where we are still developing depth, including agent orchestration, retrieval pipelines, LLM evaluation, and prompt design. This is an ongoing platform effort, not a one-time project. The work will evolve over time as we expand capabilities, integrate new tools, and improve system quality. Our core systems are built on a Microsoft/.NET stack. As we develop our agentic AI capabilities, we expect to leverage cloud-based tooling (likely within Azure) where appropriate. We are pragmatic about using the right tools for the problem, while maintaining alignment with our existing platform and operational requirements. At Midwest Family we value practical experience, clear thinking, and the ability to apply these skills in a real-world environment.

Requirements

College degree or Programming Certification (preferred)
5+ years of equivalent work experience
Experience building or working with AI/LLM-based systems in a production or near-production environment
Strong understanding of agent-based patterns (ReAct, tool usage, etc.)
Experience with retrieval-based systems (RAG), vector stores, or search pipelines
Familiarity with evaluating non-deterministic system outputs
Experience improving system quality through testing and iteration
Strong software engineering skills required. Experience with Python, TypeScript, or C# preferred
Ability to work across technologies and integrate AI solutions into an existing enterprise environment

Nice To Haves

Experience working with multiple LLM providers
Familiarity with emerging tools and frameworks in the agentic AI space
Experience integrating with enterprise systems and data sources
Experience mentoring or supporting other developers

Responsibilities

Platform Architecture & Evolution: Provide input into AI-related architecture decisions. Evaluate tools, models, and approaches, and make practical recommendations based on trade-offs and outcomes.
Agent Lifecycle Management: Design, build, and maintain agents. Define prompts and configurations, support routing decisions, and monitor performance in production.
RAG Pipeline Engineering: Improve retrieval quality by working with chunking strategies, embeddings, search methods, and metadata filtering. Support onboarding of new knowledge sources.
Quality & Hallucination Reduction: Help reduce incorrect or unsupported outputs through system design, testing, and monitoring. Apply practical approaches across prompts, retrieval, and model behavior.
Evaluation & Testing: Develop and maintain evaluation approaches to measure output quality, including relevance and accuracy. Support regression testing and identify failure patterns.
Prompt Engineering: Create and maintain prompts for different agents. Diagnose unexpected behavior and develop reusable patterns where appropriate.
Tool Integration: Support integration with internal systems and APIs. Help manage tool access and ensure reliable execution and error handling.
Observability & Operations: Contribute to logging, tracing, and monitoring. Support analysis of system behavior, cost, and performance in production.
Governance & Configuration: Work with configuration across agents, prompts, and models. Ensure changes are version-controlled and reviewed.
Developer Support: Work with other developers to share knowledge, review designs, and establish consistent approaches.
Research & Evaluation: Stay current with new models and tools. Evaluate them in a structured way and recommend adoption where appropriate.