AI Architect (Voice AI)

About the role

We are looking for an AI Architect to lead the production‑grade voice copilot for a leading US veterinary hospice network. The solution listens to live calls, extracts appointment and clinical data in real time, and updates the scheduling system via a Chrome extension. You will own the end‑to‑end technical architecture from validated proof‑of‑concept to full production.

Key responsibilities

• Design and implement the full pipeline: streaming speech‑to‑text, LLM‑based field extraction, Chrome‑extension delivery, and AWS infrastructure.
• Reduce latency (P95) from ~6 seconds to ~2 seconds and eliminate post‑processing delays.
• Run model A/B tests (e.g., Claude Haiku vs GPT Luna) with golden‑set evaluation for name and email accuracy.
• Monitor cost, accuracy, and concurrency (5‑10+ simultaneous calls) using Langfuse traces and custom dashboards.
• Harden the system for production: data isolation, monitoring, alerting, safe rollback, and demo fallback.
• Collaborate with client stakeholders, present concrete performance metrics, and keep scope aligned with the SOW.
• Lead the AI Engineer and pod, set tasks, review output, and ensure knowledge transfer.

Required profile

• Senior‑level experience designing and delivering real‑time AI services in production.
• Proven ability to work with meticulous clients and manage scope.
• Strong communication skills for technical discussions and stakeholder alignment.

Required skills

• AWS infrastructure (compute, networking, monitoring).
• Streaming speech‑to‑text pipelines.
• Large language model (LLM) integration and field extraction.
• Chrome‑extension development and deployment.
• Amazon SES for automated email briefings.
• Model A/B testing and evaluation (Claude Haiku, GPT Luna).
• Langfuse for tracing and cost/accuracy dashboards.
• Production‑grade monitoring, alerting, and rollback mechanisms.

What we offer

• Opportunity to shape a cutting‑edge voice AI product in the veterinary care sector.
• Collaboration with a private‑equity backed portfolio of AI projects.
• Potential for role expansion as production scales (up to 1 FTE).