- AI Feature Blueprinting: Create technical blueprints for AI features, including architecture, data needs, model strategy, evaluation approach, and delivery risks.
- Model Selection: Define patterns for identifying which LLM model should be used for each call, and when to use NLP/NLU instead.
- Harness Architecture: Shape reusable patterns for agent loops, tool/function calling, context construction, memory, structured outputs, retries, and failure handling.
- Prompt Engineering: Set standards for prompt design, versioning, review, experimentation, and regression testing as first-class engineering assets.
- Evaluation: Design eval frameworks, golden datasets, regression suites, error analysis, and LLM-as-judge patterns to measure quality and guide decisions.
- RAG & Knowledge Patterns: Guide document parsing, chunking, embeddings, retrieval, reranking, prompt assembly, and grounded response patterns.
- Mentorship & Enablement: Coach P2/P3 AI engineers, review technical designs, create examples, and help junior engineers present architecture findings clearly.
- Cross-Guidewire Collaboration: Partner with Product, AI CoE, PS, customer teams, and SI partners.

KPIs & Success Metrics

- AI feature blueprints are explicit, reusable, and actionable for Strike Teams.
- Model and architecture choices are backed by evals, analysis, and documented trade-offs.
- Prompt, harness, RAG, and evaluation patterns are adopted consistently by P2/P3 engineers.
- Delivered AI features meet project bars for accuracy, groundedness, reliability, latency, and cost.
- Junior and mid-level engineers can implement patterns and explain the results confidently.

Key Skills & Experience

- Education: Bachelor's or Master's degree in Computer Science, Data Science, Statistics, Machine Learning, Engineering, or a related technical field.
- Experience: 8+ years of experience in AI, machine learning, data science, software engineering, or solution architecture.
- AI Architecture:



Strong ability to translate ambiguous product or delivery goals into practical AI system designs.
- AI Fundamentals: Deep understanding of LLMs, prompts, embeddings, retrieval, structured outputs, tool use, agentic workflows, and model failure modes.
- NLP/NLU Judgment: Able to recognize when classical NLP/NLU, rules-based approaches, lightweight ML, or structured extraction are better than generative AI.
- Technical Proficiency:
- Strong Python/Java skills and ability to read, review, and contribute to production-quality code.
- Experience designing LLM harnesses, including orchestration, tool/function calling, context construction, retries, and fallback behavior.
- Strong prompt engineering practice, including structured iteration, versioning, evaluation, and regression testing.
- Working knowledge of RAG patterns, embeddings, retrieval evaluation, and vector search concepts.
- Comfortable designing evaluation approaches using golden sets, regression suites, error taxonomies, and human review loops.

- Engineering Excellence: Familiar with Git, code review, testing, CI/CD concepts, cloud-based delivery, and production engineering practices.
- Communication & Mentorship: Able to explain complex AI trade-offs clearly and coach less experienced engineers without direct reporting authority.

Preferred Skills & Experience

- Agentic AI Systems: Experience with agentic frameworks or custom agent harnesses in production.
- Evals: Experience building structured eval harnesses, regression suites, LLM-as-judge patterns, or production feedback loops.
- Cloud: Hands-on AWS experience with services such as Bedrock, SageMaker, OpenSearch, or related cloud-native AI/data services.
- Fine-Tuning: Experience evaluating or implementing model fine-tuning, distillation, or domain adaptation.
- Guidewire Knowledge: Familiarity with Guidewire products or the insurance domain is a plus.
- Enterprise AI Delivery: Prior experience working on document-heavy, regulated, insurance, finance, or professional services datasets.

📌 AI Architect (Bengaluru)
🏢 Guidewire Software
📍 Bengaluru