Remote AI Voice Engineer — Conversation Design & QA

CoreReachAI, Inc.; United States; Remote; USD 120,000 - 180,000; Full time

Benefits offered by this job

Equity

Fully remote

Health insurance

Dental insurance

Vision insurance

PTO & holidays

Home office stipend

Ownership of agent behavior

Job summary

CoreReachAI, Inc. is seeking an AI Voice Engineer for a fully remote role open to candidates in the United States. You will own how our voice agents listen, respond, escalate, and are validated on every call.

Design conversation logic, define emergency thresholds, and build guardrails to ensure fair, accurate responses. You will also develop evaluation tooling and optimize prompts and model settings for natural conversations.

Qualifications

• 3+ years building software, including production products powered by LLMs.
• Strong Python or TypeScript skills.
• Hands-on experience with prompt design, evaluation and structured output from LLMs.
• Familiarity with speech-to-text and text-to-speech systems.
• Good judgment about edge cases, and the habit of testing for them.
• Comfort working across the full stack when needed.

Responsibilities

• Design and improve the conversation logic behind our voice agents, from greeting to handoff.
• Build the rules that decide what counts as an emergency and who gets called.
• Keep the agent's answers accurate and fair-housing safe, with guardrails that hold on every call.
• Build evaluation tooling and test suites that catch problems before customers hear them.
• Tune prompts, models and settings for faster, more natural conversations.
• Monitor call quality in production and debug issues across the voice, telephony and model layers.
• Work with product and customer success to turn customer requests into agent behavior.

Skills

Python

TypeScript

Prompt design

Speech-to-text

Text-to-speech

Tools

SIP/WebRTC

Job description

CoreReachAI, Inc. is seeking an AI Voice Engineer for a fully remote role open to candidates in the United States. You will own how our voice agents listen, respond, escalate, and are validated on every call.

Design conversation logic, define emergency thresholds, and build guardrails to ensure fair, accurate responses. You will also develop evaluation tooling and optimize prompts and model settings for natural conversations.