Description

Required Skills & Experience 5+ years of Strong Software Engineering (Python/NodeJS), system design and production service experience. 2+ years of Experience with LLMs, prompt engineering, and agent frameworks. 2+ years of Experience Practical experience implementing RAG: embeddings, vector DBs and retrieval tuning. I 2+ years of Experience with LangChain patterns and with toolchain telemetry (Langfuse or similar) for prompt/model traceability. 5+ years of Experience with Kubernetes, Docker, CI/CD and infrastructure-as-code experience. 2+ years of Experience with Practical experience with Google Cloud Platform services 2+ years of Experience with Observability, testing, and security best practices for distributed systems. 2+ years of Experience with evaluating and mitigating retrieval/augmentation failures, hallucinations, and leakage risks in RAG systems. Familiarity with vendor and open-source vector stores and embedding providers. Familiarity with CI/CD pipelines (Jenkins, GitHub Actions, GitLab Cl, or ArgoCD). Core Responsibilities Implement and maintain MCP server and agent code, APIs, and SDKs for model access and agent orchestration. Design agent behavior, workflows and safety guards for agentic Al systems. Create, test and iterate prompt templates, evaluation harnesses and grounding/chain-of-thought strategies. Integrate LLMs and model providers (self-hosted and cloud APIs) with unified adapters and telemetry. Build developer tooling: CLI, local runner, simulators, and debugging tools for agents and prompts. Containerize services (Docker), manage orchestration (Kubernetes/GKE), and optimize nodes, autoscaling and resource requests. Ensure observability: logging, metrics, traces, dashboards, alerting and SLOs for model infra and agents. Create runbooks, playbooks and incident response procedures; reduce MTTR and perform postmortems. Design and maintain RAG workflows: document chunking, embeddings, vector indexing, retrieval strategies, re-ranking and context injection. Integrate and instrument LangChain for composable chains, agents and tooling; use Langfuse (or equivalent tracing) to capture prompts, model calls, RAG traces and evaluation telemetry

Key Skills

Education

Any Gradute