Insurance AI Architect

Coforge; Dadri; On-site; INR 3,000,000 - 6,000,000; Full time

Job summary

Coforge is seeking an Insurance AI Architect with 12-18 years of experience to lead AI/ML initiatives for underwriting, claims, and pricing within our insurance technology stack.

Focus areas include LLM/GenAI, MLOps, cloud platforms (Azure/AWS/GCP), Document AI, and platform integration with Duck Creek and Guidewire to drive scalable insurance solutions.

Job description

Experience: 12-18 Years

We at Coforge are hiring for Insurance AI Architect

AI/ML Technical Stack

• LLM & GenAI frameworks: LangChain, LlamaIndex, Semantic Kernel; prompt engineering, RAG architecture design, and agentic frameworks (CrewAI, or similar) for multi-step underwriting/claims workflows
• Foundation model platforms: Practical experience with Azure OpenAI, AWS Bedrock, Google Vertex AI, and Anthropic Claude/API; model selection criteria (cost, latency, accuracy) for insurance-specific tasks
• Traditional ML/predictive modeling: Regression, gradient boosting (XGBoost/LightGBM) for risk scoring and pricing models; familiarity with actuarial modeling concepts as they intersect with ML
• Document AI/IDP: OCR and intelligent document processing (Azure Form Recognizer, AWS Textract) for submission intake, ACORD forms, loss runs, and SOV (statement of values) processing a common Commercial/Specialty use case
• MLOps & deployment: Model versioning, monitoring, and governance tooling (MLflow, Azure ML, SageMaker); CI/CD for AI pipelines integrated with platform release cycles
• Microsoft AI Platform Expertise: Copilot Studio, Foundry and Microsoft ecosystem integration
• Platform decision-making: Judgment on when to use Copilot Studio (low-code, business-user-facing agents) vs. AI Foundry (custom, code-first, higher-control agentic solutions) vs. open-source/other cloud AI stacks, based on client maturity, governance needs, and use case complexity
• Cloud platforms: Azure, AWS, or GCP; landing zone design, data lake/lakehouse architecture for insurance data
• Data engineering: ETL/ELT pipelines pulling from Duck Creek/Guidewire data models into analytics or AI-ready stores; understanding of insurance data standards (ACORD, ISO)
• API & microservices design: RESTful API design, event streaming (Kafka), and how to expose AI capabilities as composable services consumable by both platforms

Insurance Industry & Platform Expertise

• Duck Creek Technologies: Working knowledge of Duck Creek OnDemand, Duck Creek Policy, Claims, Billing, and Rating modules; understanding of Duck Creek's Author/Manuscript configuration model, Duck Creek Data Insights, and Duck Creek's low-code/no-code extensibility framework
• Optional - Guidewire InsuranceSuite: Basic knowledge of PolicyCenter, ClaimCenter, and BillingCenter architecture, including Gosu scripting, PC/CC/BC data model, Integration Gateway, Cloud API, and Guidewire Cloud Platform (GWCP); familiarity with UI framework and Guidewire Marketplace apps
• Platform integration patterns: Experience with how AI services plug into both platforms via REST/SOAP APIs, event-driven architectures (Guidewire Cloud Events, Duck Creek's integration points), webhooks, and middleware (MuleSoft, Boomi, Azure Integration Services)
• Understanding of P&C Commercial & Specialty product structures: Rating algorithms, forms/endorsements, exposure bases, and how these map into platform data models

Governance & Compliance

• Working knowledge of model risk management, explainability techniques
• Familiarity with SOC 2, data privacy (state insurance data regulations), and responsible AI frameworks as applied to carrier environments