GenAI Solution Architect - Hartford, CT (Onsite)

Role Name: Gen AI Solution Architect

Location: Hartford, CT (Onsite)

Start date: 12  Months

Role summary

Own the end-to-end architecture of GenAI and agentic solutions on Google Cloud for insurance clients. You shape solutions in pre-sales, set the technical direction for delivery teams, and make sure what ships is secure, governed, scalable and cost-efficient.

Key responsibilities

Define reference architectures for RAG, document intelligence and multi-agent systems on Gemini Enterprise Agent Platform (Vertex AI).

Choose models (Gemini family, Model Garden, open models) based on accuracy, latency and cost; set model-routing and fallback patterns.

Design agent architectures with ADK, Agent Engine, A2A and MCP; define tool, memory and session strategy.

Design the data foundation: BigQuery, Cloud Storage, AlloyDB, Vector Search / Vertex AI Search, Document AI, Dataflow and Pub/Sub.

Set integration patterns with policy admin, claims and underwriting systems (for example Guidewire, Duck Creek) through Apigee and event-driven APIs.

Own security and compliance: VPC Service Controls, CMEK, IAM, Sensitive Data Protection, audit logging, data residency.

Set up LLMOps: evaluation frameworks, prompt and model versioning, monitoring, drift and hallucination tracking, human-in-the-loop review.

Lead Responsible AI reviews: explainability, bias testing, model risk documentation for insurance regulators.

Estimate effort and cloud cost (tokens, compute, storage) for proposals and SOWs; build POCs that win deals.

Present to client architecture boards and CXOs; mentor GenAI leads and engineers across programs.

Required skills

Deep, hands-on Google Cloud Platform: Gemini Enterprise Agent Platform (Vertex AI), Gemini API, Agent Builder, Cloud Run, GKE, BigQuery, Apigee, IAM and networking.

Proven production RAG and agentic systems at enterprise scale.

Python and solution prototyping; comfortable reviewing engineers' code.

LLM evaluation, guardrails and cost optimization (context caching, batch prediction, model tiering).

Infrastructure as code (Terraform) and CI/CD for ML/GenAI.

Strong stakeholder management, solutioning and documentation (ADRs, HLD/LLD).

Nice to have

Insurance or financial services domain: underwriting, submissions intake, claims, COI, E&S.

Multi-cloud GenAI exposure (Azure OpenAI, AWS Bedrock) for comparative solutioning.

Gemini Enterprise (formerly Agentspace) rollout for employee-facing agents.

Model tuning and distillation; Looker semantic layer for analytics agents.

Certifications (preferred)

Google Cloud Professional Cloud Architect (required or within 3 months)

Google Cloud Professional Machine Learning Engineer

Google Cloud Generative AI Leader

Best Regards,

Chetna

Truth Lies in  Heart