Experience : 10+ yrs Work Location : Remote Define orchestrator + domain‑agent blueprint: request routing, contracts, memory strategy, tool usage, and safety/guardrails. Establish scalability patterns on GKE (HPA, node pools, autoscaling, pod disruption budgets) and low‑latency streaming/messaging. Integrate Gemini / ADK cleanly with existing services; plan migration off bespoke REST where appropriate. Set observability (traces, metrics), evaluation harnesses, A/B testing, and rollback strategies for LLM/agent changes. Partner with security for data boundaries, PII handling, prompt‑injection mitigations, and scoped permissions. Create decision records (ADRs), risk registers, and a phased roadmap aligned to cost, latency, and reliability targets. 8–10+ years in distributed systems; 2–3+ years building LLM/agentic systems at enterprise scale. Hands‑on GCP (GKE, Cloud Run, Pub/Sub, Cloud Logging/Trace/Monitoring, Secret Manager), Gemini (incl. ADK/SDK), and SSE streaming. Track record delivering reference architectures and leading production launches. RAG patterns (vector stores, chunking strategies, retrieval and re‑ranking policies); LangChain/LlamaIndex familiarity. Cost optimization (autoscaling policies, model selection, caching). Experience turning a single assistant into an orchestrator of domain agents. if you are interested, please share your updated resume to
Job Title
Immediate opening for LLM Systems Architect — Multi‑Agent Enterprise Architecture (GCP/Gemini)