AI Solutions Architect — Agentic AI Platforms
San Francisco, California, United States of America
Full Time
2 weeks ago
H1B Sponsor Likely
Key skills
AIClaudeRAGLangChainAgenticMLOpsLangGraphQdrantKafkaPlaywrightKubernetesPostgreSQLPrometheusGrafanaOpenTelemetrySaaSBudgetingCustomer SuccessSales
About this role
SaaS Technologies is seeking an AI Solutions Architect to lead the design and delivery of an enterprise Agentic AI platform. The role involves owning the technical vision for multi-agent systems and driving the integration of various tools and microservices to enable enterprises to operate domain-specific agents at scale.
Responsibilities:
- Agentic AI Architecture: Own end-to-end design of multi-agent systems using LangChain, LangGraph, and Model Context Protocol (MCP) — including planner-executor patterns, sub-agent hierarchies, tool routing, retries, and cost-aware token budgeting
- RAG & Knowledge Systems: Architect production-grade RAG pipelines with vector databases (pgvector, Qdrant), hybrid retrieval, re-ranking, and document-aware chunking to ground agents in enterprise knowledge
- Solution Architecture: Design reference architectures and solution blueprints for enterprise clients across regulated and consumer-facing industries — translating business outcomes into agentic AI roadmaps and reusable accelerators
- Scalable Microservices: Build event-driven microservices on Kafka, polyglot data layers with PostgreSQL and vector DBs, and Kubernetes-based deployment topologies for high-throughput inference workloads
- MLOps & Model Lifecycle: Establish practices spanning training, fine-tuning, prompt and config versioning, structured evaluations against golden datasets, drift detection, and automated rollback when output quality degrades
- Traceability & Observability: Instrument agent reasoning traces, tool-call audit trails, token spend, and quality signals with Prometheus, Grafana, and OpenTelemetry — enabling policy enforcement and human-in-the-loop oversight
- Reusable Engineering Standards: Codify AI engineering patterns (RAG retrievers, agent loops, eval harnesses, traceability spans) into reusable skills, sub-agents, and platform components consumed across multiple product lines
- Rapid Engineering in Agentic Development Lifecycle: Roll out AI-led developer tools and sub-agents (Claude Code, Playwright MCP) across planning, code generation, code review, test authoring, and release validation — accelerating delivery while standardizing quality
- Presales & Client Engagement: Partner with sales, presales, and customer success on enterprise pursuits — authoring solution designs, leading technical workshops, and shaping agentic AI roadmaps for prospects and existing clients