Agentic AI Engineer

MAPFRE is hiring an Agentic AI Engineer to design, build, and ship production-grade AI agents that help improve high-impact insurance workflows, including underwriting and claims. This hybrid role in Webster, MA brings a hands-on engineering focus with end-to-end ownership, from agent architecture and backend services to evaluation, guardrails, and scaled deployment.

You will create AI agents and multi-step agentic workflows that turn complex insurance processes into reusable engineering components and patterns. You’ll translate solution and business designs defined with the AI architect and product owner into working agents, balancing trade-offs across deterministic rules, traditional ML, plain GenAI, and autonomous agent approaches.

• Design, build, and deploy AI agents and multi-step agentic workflows for underwriting and claims, creating reusable components and engineering patterns to accelerate future solutions
• Advise on technical feasibility, cost, and reliability trade-offs when converting business and solution designs into agent implementations
• Implement agents as backend services (for example,
  FastAPI
  ) with stable API contracts, secure authentication, resilient error handling, and automated testing
• Integrate agents with enterprise systems, data sources, and APIs using tools and MCP servers
• Own agent quality through success criteria, test datasets, automated evaluations, failure diagnosis, and targeted improvements
• Implement guardrails including human-in-the-loop checkpoints, action limits, observability, monitoring, performance and reliability controls, cost management, and audit logging aligned with AI governance and regulatory requirements
• Prototype and validate solutions in collaboration with business and technical teams, including lightweight MVP interfaces (for example,
  Streamlit
  ), then deploy and scale using cloud, containerization, and CI/CD
• Measure technical performance and business value, and communicate results and limitations to technical and non-technical stakeholders

• Bachelor’s degree in Computer Science, Engineering, or a related quantitative field (or Master’s degree), with either 7+ years of software, machine learning, or data engineering experience or a Master’s with 5+ years of relevant experience
• 2+ years of hands-on experience building LLM-based applications, including at least one agentic system (tool calling, multi-step workflows, etc.) deployed to production and used by real users
• Strong Python skills and solid software engineering fundamentals, including API and tool integration, automated testing, error handling, version control, and debugging of non-deterministic systems
• Hands-on experience with at least one agent framework or SDK (
  Lang Graph strongly preferred
  ) and major model APIs or managed AI platforms such as Amazon Bedrock
  , building production AI agents that integrate with tools, APIs, and enterprise systems
• Experience evaluating LLM applications or agents, including test datasets, metrics, regression testing, and appropriate use of model-based evaluators
• Understanding of agent reliability and security risks, with pragmatic judgment, strong problem-solving skills, and clear communication with technical and business stakeholders
• Ability to solve ambiguous, complex problems with minimal direction and high autonomy, with a results-driven mindset and persistence to find root causes
• Strong communication skills to explain agent behavior and limitations to non-technical stakeholders

• Python, FastAPI, Streamlit, Lang Graph, Amazon Bedrock, MCP

• Paid Time Off (PTO) - flexible PTO program
• 401(k) & Profit Sharing (company matching contributions up to certain amounts)
• Medical, Dental and Vision Coverage
• STD,…