Job description

Senior Software Engineer (Voice AI) Location: New York City | In-office THE MISSION “Leading healthcare into the agentic era.” Every day, patients call their clinic to schedule an appointment, ask a question, or request a prescription refill. Behind each conversation is a busy healthcare team coordinating people, information, and systems. Taxo builds Voice AI agents that take on this administrative work. Our agents connect with EHR and practice management systems to help teams handle patient intake, complex scheduling, and refill requests, with reliable workflows and clear handoffs to people. THE OPPORTUNITY: OWN THE VOICE EXPERIENCE As our Senior Software Engineer (Voice AI), you’ll report to the CTO and own defined voice-runtime services that make a voice agent useful in a real healthcare conversation. You’ll build and operate speech, turn-taking, interruption handling, tool orchestration, observability, and production fixes for the services you own, while contributing to broader platform decisions with the CTO and engineering team. This is a hands-on role with direct influence on how the product works. WHAT YOU’LL DO - Build the Core: Build and improve real-time voice agents for healthcare administrative workflows, including intake, scheduling, and refill requests. - Connect the Pipeline: Develop the conversation pipeline across speech recognition, language models, speech generation, and telephony integrations. - Get the Conversation Right: Improve turn-taking, interruption handling, response latency, and recovery when a conversation or external system goes off track. - Integrate Deep: Build voice-agent orchestration and tool calls on shared backend APIs and EHR connectors, partnering with the engineers who own those services and integrations. - Measure Voice Performance: Own voice-component tests and production instrumentation for latency, turn-taking, and tool execution. Partner with the quality engineering team on shared evaluation tooling and cross-product regression coverage. - Own Voice Production: Own voice-service observability, on-call response, diagnosis, and fixes. Use quality findings and rollout feedback to improve live voice behavior. THE STACK We move fast with a modern toolkit: - Frontend: React, Next.js - Backend: Node.js (TypeScript) - Data & Infrastructure: PostgreSQL, AWS - Intelligence: LLMs and Voice AI integration WHO YOU ARE - The Builder: Hands-on experience shipping and operating real-time voice AI systems in production, including turn-taking, interruption handling, latency, and recovery from failures. - The AI Engineer: Strong software engineering skills and practical experience with LLMs, agent frameworks, tool calling, and external APIs. - The Problem Solver: Practical experience balancing latency, accuracy, cost, and voice conversation quality in live systems. - The Architect: The ability to debug distributed systems and take ownership of reliability, observability, and testing. - The Product Partner: Product judgment, clear communication, and a track record of moving from an ambiguous problem to a working solution. - The Owner: Careful judgment around sensitive data, access controls, and safe escalation to a human. - Bonus: Experience with streaming audio, SIP/VoIP, healthcare workflows, EHR integrations, or privacy-sensitive systems. WHY TAXO Healthcare teams should be able to spend more time on patients and less time on administration. Help us build the products and company that make that possible. - 🧠 Frontier AI Work: Work on voice agents, real-time inference, tool use, and the hard problems of conversational quality. - 🏥 Impact That Matters: Build systems that reduce administrative work and make it easier for patients to reach their care teams. - 🛠️ Real Ownership: Take a problem from the first prototype through release, measurement, and improvement. COMPENSATION Base salary: $180,000–$220,000 USD annually Equity: 0.20%–0.50% Benefits: Health coverage HOW WE WORK Taxo is a small, fast, high-conviction team. We care deeply about the problem, move with urgency, and take ownership of the work. We collaborate closely, make practical decisions, and keep improving through what we learn from customers. If you want to build voice AI that works in the real world of healthcare, we’d love to hear from you.

Requirements

Must-haves

Senior Software Engineer (Voice AI)
Location: New York City | In-office
“Leading healthcare into the agentic era.”
Every day, patients call their clinic to schedule an appointment, ask a question, or request a prescription refill. Behind each conversation is a busy healthcare team coordinating people, information, and systems.
Build the Core: Build and improve real-time voice agents for healthcare administrative workflows, including intake, scheduling, and refill requests.
Connect the Pipeline: Develop the conversation pipeline across speech recognition, language models, speech generation, and telephony integrations.
Get the Conversation Right: Improve turn-taking, interruption handling, response latency, and recovery when a conversation or external system goes off track.
Integrate Deep: Build voice-agent orchestration and tool calls on shared backend APIs and EHR connectors, partnering with the engineers who own those services and integrations.
Own Voice Production: Own voice-service observability, on-call response, diagnosis, and fixes. Use quality findings and rollout feedback to improve live voice behavior.
Frontend: React, Next.js
Backend: Node.js (TypeScript)
Data & Infrastructure: PostgreSQL, AWS
Intelligence: LLMs and Voice AI integration
The Builder: Hands-on experience shipping and operating real-time voice AI systems in production, including turn-taking, interruption handling, latency, and recovery from failures.
The AI Engineer: Strong software engineering skills and practical experience with LLMs, agent frameworks, tool calling, and external APIs.
The Problem Solver: Practical experience balancing latency, accuracy, cost, and voice conversation quality in live systems.
The Architect: The ability to debug distributed systems and take ownership of reliability, observability, and testing.
The Product Partner: Product judgment, clear communication, and a track record of moving from an ambiguous problem to a working solution.
The Owner: Careful judgment around sensitive data, access controls, and safe escalation to a human.
Healthcare teams should be able to spend more time on patients and less time on administration. Help us build the products and company that make that possible.
Frontier AI Work: Work on voice agents, real-time inference, tool use, and the hard problems of conversational quality.
Impact That Matters: Build systems that reduce administrative work and make it easier for patients to reach their care teams.
Real Ownership: Take a problem from the first prototype through release, measurement, and improvement.

Nice-to-haves

Bonus: Experience with streaming audio, SIP/VoIP, healthcare workflows, EHR integrations, or privacy-sensitive systems.

Application prep

Use the role’s language and application details while tailoring your materials.

Keywords to mirror

aireactnext.jsnode.jstypescriptaws

Platform: ashby
Effort: quick