Role & responsibilities :

Build the agentic AI tower from zero : agents in the execution path of customer care, retention and field operations orchestration, tool integration, eval harnesses, guardrails, cost control and hire the developer team.

Preferred candidate profile :

Strong backend/ML engineer - LLM application builder - led a small AI team or platform (8 - 12 years overall). The spine is software engineering; the last 2 - 3 years are production LLM systems.

Mandatory requirements :

- Shipped LLM-powered systems to PRODUCTION with real users - demos and POCs do not qualify.

- Agent patterns hands-on : tool/function calling, MCP-style integration, orchestration, structured outputs.

- Built evaluation harnesses and guardrails - can explain how they KNEW the system worked and what it cost.

- Strong engineering spine : Python and/or TypeScript/Java, APIs, queues, databases - an engineer first.

- Has hired and led engineers; built a team or practice from zero.

Nice to have :

- Claude/Anthropic API depth (or equivalent frontier-model platform mastery).

- Contact-center / voice-bot / WhatsApp-channel AI; RAG over messy enterprise data.

- LLM cost-and-latency engineering at scale; MLOps exposure.

Tech Stack :

Frontier LLM APIs (Claude-class), MCP / tool calling, RAG + vector stores, eval frameworks, Python + TypeScript, Kafka/REST plumbing, Kubernetes.