AI Voice Systems Engineer for Elder Care Coordination

Client’s mission is to make elder caregiving easier through AI-powered care coordination. We help older adults remain independent by providing practical solutions that improve their daily lives while giving family caregivers the support they need. Our voice system provides personalized experiences to elder customers by assisting them in arranging rides, shopping, tracking calendars, and getting help for in-home maintenance issues—all through voice calls only. It also assists caregivers in performing multiple types of check-ins with their loved ones. The right candidate will bridge the gap between cutting-edge technology and real-world business needs.

Requirements

Key technologies:

• Experience working with containerized AI workloads on AWS.
• Key components: Cursor, OpenAI, Twilio, 11 Labs, gpt-4.0-mini, S3, Amazon Connect, KVS (voice), Swagger, OpenAI tools.
• Database: Postgres.
• Ticket management system: Linear.
• Programming languages: Proficiency in TypeScript/JavaScript for real-time web/edge integrations.
• Natural language processing libraries and speech processing toolkits (examples: knowledge of deep learning libraries such as PyTorch, TensorFlow).
• Speech recognition APIs (example: Amazon Transcribe).
• Voice synthesis engine (examples: Amazon Polly, Text-to-Speech).
• Knowledge of version control systems (GitHub), CI/CD pipelines, and containerization (Docker) is highly valued.
• Networking & Protocols: Understanding of real-time communication protocols, WebSockets, gRPC, and WebRTC.

AI Requirements:

• Realtime/streaming LLM sessions: latency, interruption handling, turn-taking, and graceful degradation when the model or connection drops.
• Prompt and system-message design for a constrained, task-specific voice agent (not open-ended chat).
• Sentiment/outcome extraction from a completed call and how it's validated.

Responsibilities

• Audit the existing Twilio ↔ Realtime-AI call bridge (media streams, ConversationRelay) and the two-way contracts with the AWS-hosted platform (inbound invokes, outbound callbacks, retries, reconciliation).
• Assess production-readiness of the voice service — health checks, capacity limits, alerting, incident response — and identify gaps.
• Review the AI/session logic driving calls, along with related routes, background jobs, and schema.
• Deliver a findings report and a prioritized implementation roadmap.
• Provide part-time advisory support through implementation: review architectural/technical decisions, guide webhook-contract and telephony-integration design, and consult on production operability practices.