Full Time
950
40
Jul 26, 2026
About us
We run AI voice agents that call US insurance companies on behalf of healthcare clinics to verify patient benefits. The AI navigates phone trees, talks to real reps, and pulls back the coverage details the clinic needs. It works well, but voice agents live and die on the details, and we need someone watching them every day and making them sharper.
What the job actually is
You will sit and listen to real calls our AI makes. For every call you listen to, you do three things:
Take notes on where the call could be tighter, faster, or clearer.
Turn those notes into real improvements to how our AI behaves on the next call.
When something breaks badly on a live call, you jump into the code, figure out what went wrong, and get a fix live fast.
That last part is the one that matters most. If a call has a critical error, I need someone who can pull up the codebase, understand it, prompt an AI to help fix it, and ship the fix within the hour. Not tomorrow. Not after a meeting. Within the hour.
This is a monitoring plus building role. Think of it as being the person who keeps the AI honest and keeps making it smarter.
Who I am looking for
I am looking for a strong vibe coder more than a traditional heads down engineer. Someone who lives in tools like Claude Code or Codex, who can move through a codebase quickly with AI help, and who ships. You should be comfortable reading code you did not write and changing it safely.
You must have real experience with:
Building AI voice agents (Twilio or similar). You have made AI place or take real phone calls before.
Prompt architecture. You know how to shape an agent's behavior through prompts and you understand why small wording changes matter.
Navigating IVRs, DTMF tones, hold music, phone trees, all the messy real world stuff that comes with automating phone calls.
Strongly preferred:
Experience in US healthcare or insurance. You know what an EOB is, you understand benefits verification, or you have worked around medical billing.
Excellent written and spoken English. This is not negotiable. You will be reading transcripts, writing clear notes, and possibly speaking, so your English has to be genuinely strong.
Brownie points (tell me if you have these):
AWS, especially CloudWatch and S3.
Any experience running production systems and reading logs to find problems.
Hours
This is on US Eastern time (EST). You need to be available and alert during US business hours, which is a night shift in the Philippines. Please only apply if you can genuinely commit to those hours.
Tools
You need your own subscription to an AI coding CLI, either Claude Code or Codex or something equivalent. We will help with some of the budget for this once you are on board, but you should already be using one of these daily and be able to show me that.
How to apply
Do not send me a generic resume and nothing else. I will ignore it. Instead, send me a short video of yourself where you:
Speak (in English) about AI voice agents you have built before and what they did.
Tell me about a time you fixed something broken fast, and how fast you actually work.
Show me your screen while you use Claude Code, Codex, or whatever CLI you use. I want to see you actually moving through code with AI, not just talking about it.
If you have healthcare or insurance experience, mention it clearly in the video.
The people who show me real work and real speed will go to the top of the pile. I care far more about what you can actually do than about a polished resume.