Senior AI Engineer

Guidewire Software•Mississauga, ON

10d

About The Position

The Senior AI Engineer at Guidewire is responsible for defining and driving the architecture of complex agentic flows, ensuring reliability, steerability, and seamless integration within the Guidewire ecosystem. This role involves collaborating with subject matter experts to translate intricate insurance logic into agentic reasoning patterns, adhering to strict regulatory and business constraints. Key responsibilities include designing and overseeing high-fidelity prototypes for autonomous navigation of insurance workflows, establishing rigorous frameworks for evaluating non-deterministic agentic behavior, and implementing comprehensive monitoring and tracing for agentic decision-making processes. The engineer will also partner with core engineering to build specialized infrastructure for long-running agentic tasks, act as a subject matter expert in applied agentic AI research, and mentor other engineers on agentic design patterns. Guidewire is a platform for P&C insurers, combining digital, core, analytics, and AI as a cloud service, serving over 540 insurers in 40 countries. The company fosters a culture of curiosity, innovation, and responsible AI use, valuing integrity, rationality, and collegiality.

Requirements

Experience: 8+ years implementing ML/AI technologies, with at least 2 years focused on building and scaling LLM-driven applications.
Agentic Mastery: Proven track record of building and deploying agentic workflows. Deep expertise in orchestration frameworks (e.g., LangGraph).
Advanced Prompt Engineering & Tool-Use: Expertise in teaching models to use external APIs, databases, and legacy tools effectively. You should have a deep understanding of tool-calling, structured output generation, and context window management.
Domain Expertise: Demonstrated experience taking AI models from notebooks to high-availability environments within highly regulated industries (e.g., P&C Insurance, Banking, or Healthcare).
Technical Proficiency: Advanced Python skills; deep familiarity with state-of-the-art LLMs (OpenAI, Anthropic).
Cloud Native Agent Infrastructure: Strong experience with AWS or similar cloud providers, with a focus on building scalable, event-driven architectures that support asynchronous agent tasks.
Problem Solving: A creative builder mindset capable of architecting solutions in the face of the ambiguity inherent in greenfield agentic projects.
Communication: Exceptional ability to explain how and why an agent made a specific decision to executive stakeholders, maintaining transparency in autonomous systems.
Demonstrated ability to embrace AI and apply it to your current role as well as data-driven insights to drive innovation, productivity, and continuous improvement.

Nice To Haves

Experience with graph-based data structures is a plus.

Responsibilities

Orchestration Architecture: Define and drive the architecture for complex agentic flows, ensuring agents remain reliable, steerable, and seamlessly integrated into the Guidewire ecosystem.
Domain-Specific Reasoning: Work closely with subject matter experts to translate intricate insurance logic into agentic reasoning patterns (e.g., ReAct, Plan-and-Execute). You will ensure agents adhere to strict regulatory and business constraints.
Rapid Agent Prototyping: Design and oversee high-fidelity prototypes, exploring how agents can autonomously navigate insurance workflows to validate business relevance.
Agentic Reliability & Evals: Set the standard for evaluating non-deterministic agentic behavior. You will build rigorous frameworks to test for tool-calling accuracy, reasoning consistency, and safety guardrails to ensure production-readiness.
Agentic Observability & Tracing: Implement comprehensive monitoring and tracing for agentic decision-making processes. You will ensure full visibility into multi-step reasoning chains, state transitions, and tool-usage to facilitate debugging, auditability, and performance tuning in complex, asynchronous environments.
Engineering Collaboration: Partner with core engineering to build the specialized infrastructure required for long-running agentic tasks.
Applied Agentic Research: Act as SME on agentic AI, conducting deep research into emerging patterns like multi-agent collaboration, autonomous self-correction, and small language model (SLM) optimization for specific agentic tasks.
Mentorship & Excellence: Elevate the organization by mentoring engineers on agentic design patterns, fostering a culture of technical curiosity.