AI Agent Developer

We're building voice AI agents for healthcare, real conversations, real patients, real impact. If you've shipped agentic systems and want to work on something that matters, read on.

You'll be part of a small, fast-moving team building a production voice AI agent on Google ADK. The agent handles inbound patient calls, appointment scheduling, cancellations, and edge-case escalations, all in real time over live audio streams.

This is not prompt engineering. You'll be deep in:

— Multi-turn voice sessions over WebSocket / Twilio Media Streams

— Tool calling pipelines with asyncio concurrency

— Prompt architecture and LLM eval pipelines

— Latency debugging on live audio (VAD, turn timing, ack gaps)

• Hands-on experience in building and shipping agentic LLM systems (not just RAG or chatbots — actual agents with tool calling, state management, and multi-turn reasoning)
• Strong Python — asyncio, FastAPI, WebSocket handling
• Hands-on experience with at least one agent framework: Google ADK, LangChain, LlamaIndex, CrewAI, or similar
• Familiarity with LLM prompt design and evaluation (you understand why a prompt fails, not just that it does)
• Experience integrating external APIs as agent tools
• Comfortable debugging live systems — logs, traces, latency profiling
• Building eval/testing frameworks for LLM outputs is plus
•  Healthcare or HIPAA-adjacent domain experience is plus
• Demonstrated ability to work alongside AI coding tools (Claude Code, Cursor or similar), we are an AI-first team and expect you to use these tools fluently, not avoid them