Key skills
KubernetesPostgresPythonLLMOpenAIClaudeAnthropicRAGOpenTelemetryGitOps
About this role
Role Overview
- Design and evolve conversational agents in production at high volume, with WhatsApp as the primary channel
- Orchestrate multi-agent systems: intent routing, handoff, session management, memory and failure recovery
- Develop tools and MCP servers that connect agents to the business domain
- Instrument LLM observability (Langfuse / LangSmith / OpenTelemetry): cost, latency, tool-call rate and production quality
- Build and maintain automated evals — no prompt or model change goes live without measurement
- Integrate multiple LLM providers (OpenAI, Anthropic, Groq) with fallback strategies and cost controls
- Implement RAG pipelines (embeddings, pgvector, reindexing)
- Create and maintain Claude Code skills, subagents and workflows that automate the delivery cycle
- Ensure reliability: rate limiting, durable queues, anti-bot guardrails, error tracking
- Operate agents on Kubernetes: GitOps deployments, resources, metrics and manifests
- Collaborate with product and design during discovery
Tech Stack
- Kubernetes
- Postgres
- Python