AI Engineer - Voice

Harper•San Francisco, CA
•Onsite

About The Position

The AI Engineer will architect and build voice AI systems for an AI-powered commercial insurance brokerage. Voice agents will support growth, sales, operations, underwriting, and customer service, handling tasks from prospecting and lead qualification to long insurance applications and 24/7 policy servicing. The role starts with rapid prototyping and deployment using low-code platforms such as VAPI or Retell, then evolves toward custom infrastructure using LiveKit, Pipecat, and related systems as usage scales. The goal is to deliver reliable, accurate, experience-focused conversations with low latency, including sub-700ms pipeline latency and support for high call volumes. The position covers the full voice stack: telephony integration, real-time media infrastructure, speech-to-text and text-to-speech, conversational intelligence, orchestration, analytics, testing, and monitoring. The engineer will build outbound and inbound calling systems, workflows for insurance applications and claims, customer service agents, and tools that let non-technical teams create and adjust voice workflows. The role emphasizes rapid deployment, iteration using real conversation data, and measurable business outcomes. This is an early-stage startup role working directly with the CEO and CTO. The position requires working from the company’s San Francisco office 5.5 days per week. Relocation costs are covered.

Requirements

  • At least 5 years of software engineering experience, including at least 3 years focused deeply on voice or audio systems.
  • Production voice AI experience building and deploying systems handling at least 10,000 minutes per month.
  • Hands-on experience with low-code voice platforms such as VAPI or Retell.
  • Deep understanding of telephony and media protocols, including SIP, RTP, and WebRTC.
  • Experience with voice orchestration frameworks such as LiveKit, Pipecat, Daily, or custom-built solutions.
  • Advanced audio processing knowledge, including voice activity detection (VAD), acoustic echo cancellation (AEC), and noise suppression.
  • Proven ability to achieve sub-second latency in production voice systems.
  • Strong proficiency in Python and TypeScript/Node.js for real-time systems.
  • Experience with inbound and outbound calling at scale, including 1,000 or more concurrent calls.
  • Expertise with modern AI providers and speech technologies, such as OpenAI, Anthropic, Deepgram, and ElevenLabs.
  • Track record of shipping voice products that directly affect business metrics.
  • Strong debugging skills for complex, multi-service voice pipelines.
  • Must be based in San Francisco and work from the office 5.5 days per week; relocation assistance is provided.
  • Experience building production voice AI systems handling 100,000 or more minutes per month with real customers.
  • Ability to maintain conversation context through calls lasting 30 minutes or more.
  • Experience optimizing voice pipelines for sub-second latency while maintaining quality.
  • Experience with both modern speech-to-text and text-to-speech providers.
  • Ability to balance practical initial solutions with technical improvements as systems scale.

Responsibilities

  • Deploy production voice agents using VAPI or Retell, then transition suitable workloads to custom infrastructure using LiveKit and Pipecat.
  • Architect and optimize voice systems for low latency, conversation quality, high call volume, and concurrent calls.
  • Integrate telephony infrastructure, including Twilio, Telnyx, and enterprise SIP systems.
  • Build outbound prospecting agents that qualify leads, handle objections, and book appointments.
  • Develop lead nurturing, predictive dialing, call pacing, and qualification workflows.
  • Build insurance form-filling agents for lengthy application conversations and underwriter follow-up systems.
  • Create document collection workflows and escalation paths to human underwriters.
  • Develop policy servicing agents, claims intake systems, payment processing agents, and proactive customer outreach workflows.
  • Create tools that allow non-technical teams to create and modify voice workflows.
  • Develop conversation analytics, A/B testing frameworks, reusable voice agent templates, and monitoring for latency, accuracy, quality, and outcomes.
  • Build systems for conversation recording, quality monitoring, and compliance.
  • Develop orchestration for handoffs between voice, SMS, and email, as well as intelligent routing to specialized agents.
  • Ship voice features frequently and iterate based on real conversation data and business metrics.

Benefits

  • Relocation assistance
© 2026 Teal Labs, Inc
Privacy PolicyTerms of Service