AI Solution Architect

Sourced from justjobs.com

The AI Solution Architect designs and implements architectural components for AI agents within Blue Cross Blue Shield of Massachusetts' Bluefield framework. Working closely with Generative AI Engineers, the role converts architectural guidelines into deployable, high-quality systems and serves as a technical evangelist through internal and external tech talks.

• Architect multi-agent systems and orchestration patterns on Vertex AI Agent Builder
• Produce solution design documents, architecture decision records, and integration patterns
• Evaluate and optimize foundation models for cost, latency, accuracy, and context fit
• Design RAG pipelines using Vertex AI Search, Vector Search, and custom data stores
• Define guardrails, safety controls, and human-in-the-loop checkpoints
• Build observability into agent architectures for tracing, logging, and monitoring
• Lead proof-of-concepts to validate architectural decisions
• Mentor the team and conduct design reviews
• Collaborate with data, platform, and security teams to align AI solutions with enterprise standards

• Bachelor's degree in Computer Science, Data Science, Engineering, or equivalent experience
• 8-10+ years in solution architecture, software architecture, or senior technical roles in cloud architecture (GCP, AWS, or Azure)
• 3+ years hands-on designing and deploying production Generative AI systems including RAG, vector databases, and semantic search
• Experience designing multi-agent systems and orchestration workflows (ADK, LangGraph, CrewAI, AutoGen)
• Strong LLM fundamentals: prompt engineering, embeddings, vector search, fine-tuning, distillation
• Understanding of ML foundations, NLP, model evaluation, and LLM safety/guardrails
• Hands-on MLOps/LLMOps experience including CI/CD, data versioning, model monitoring, and governance
• Cloud-native solution design expertise for scalability, performance, privacy, and security
• Professional Cloud Architect or AI/ML Specialty certification (GCP, AWS, or equivalent)

• Knowledge of the healthcare payer domain
• Experience with multi-modal agents (vision, audio, document processing)
• Experience building evaluation frameworks for agent quality (benchmarks, LLM-as-judge)