As part of Guidewire’s Product organization, you’ll join a team focused on embedding AI and Gen. AI across our cloud platform to transform how P&C insurers operate and innovate. We work at the intersection of insurance, data, and engineering to deliver secure, scalable products that customers can trust in highly regulated environments. You’ll collaborate across Product, PDO, Professional Services, and Go-To-Market to ship AI-powered capabilities that drive measurable customer value worldwide. The Vision. At Guidewire, we aren't just adding a "chat" bubble to our software. We are reimagining the $2 trillion P&C insurance industry as AI-native. This means moving from static forms to dynamic, agentic workflows that can read a 100-page policy, analyze a legal demand, or predict a claim's trajectory in seconds. We need an AI Product Engineer who treats LL - Ms as a raw material for building magic. You are an engineer with "Product Taste"—someone who cares as much about the user's "aha!" moment as they do about the latency of a vector search. Why This Role is Unique. In most big companies, engineers are siloed. In this role, you are embedded with the Product team. You are a co-author of the roadmap. You have the massive data and customer reach of Guidewire, but the autonomy and "move fast" culture of an AI-native startup."We are looking for the engineer who gets annoyed when an AI response is technically correct but contextually useless—and then stays up late to fix the prompt architecture because they care about the product experience."Job Description. The "AI Intuition" Factor. We care less about your years of experience and more about your intuition for the non-deterministic. You understand that:Context is King: The difference between a "hallucination" and a "breakthrough" is often just how you architect the context. The UX of AI is Different: Traditional UI is about constraints; AI UI is about guidance. You know how to build interfaces that build trust. Prototyping is Research: You’d rather get out a vibe-checked prototype in two days to get user feedback than spend two months over-engineering a pipeline that nobody wants. What You’ll DoDesign Agentic Workflows: Partner with P - Ms to decompose complex insurance tasks (like subrogation or underwriting) into multi-agent systems using agent frameworks. Rapid Prototyping: Build the "Version 0" of AI features. You will be the first person to turn a PM’s "What if..." into a working demo that feels like the future. Own the "Eval" Loop: You won't just write code; you’ll define what "good" looks like. You’ll build the evaluation frameworks to measure accuracy, tone, and reliability across thousands of insurance scenarios. RAG/graph Architecture: Implement sophisticated Retrieval-Augmented Generation and graphs that can navigate the labyrinth of Guidewire’s multi-tenant data structures. Bridge the Gap: Act as the technical translator between the "Art of the Possible" (Frontier AI) and the "Reality of the Regulated" (P&C Insurance). Your Toolkit. The AI Stack: Deep experience with OpenAI/ Anthropic APIs, vector databases, and orchestration layers. The Core Stack: You are a polyglot who can move between Python (for the AI/ ML heavy lifting) and Type. Script/ Java (to integrate with the Guidewire platform). The Product Mindset: You’ve likely built your own side projects or worked at an early-stage AI startup. You understand user friction and how to solve it with LL - Ms. Problem Solving: You don’t wait for a spec. You see a manual, "boring" insurance process and immediately start mapping out the agents needed to automate it. Experience with Voice agents a plus