My initial post about the AI interview process with Mercor was a hit—it’s been viewed over 1,400 times by the Mozibox community.
Since then, I’ve accepted a contract and started working on a project. A number of physicians have reached out asking about my experience, so I wanted to share a quick update here. As before, this isn’t an endorsement—just sharing my personal experience.
So far, so good.
When I joined the project Slack channel, several other physicians had already been involved for a few weeks. It’s a collaborative space. People are actively sharing ideas, and the energy is fun and supportive. It’s not just solo work in a vacuum. You're part of a community of doctors, each contributing from their respective specialties.
The work has been intellectually engaging. I recently submitted my first AI prompt related to occupational medicine. Writing challenging prompts has taken some adjustment, but the reviewer provided clear feedback and highlighted common areas for improvement, which helped me get up to speed.
Here’s a basic rundown of what the project involves:
You can think of the assignment as helping to give the AI model more context about your medical field. Large language models know a lot. They’ve read medical literature, textbooks, and online content but they often lack the real-world clinical judgment and nuance that physicians bring. We start by creating prompts—realistic, made-up clinical scenarios that reflect common situations in our specialties. These serve as test cases for the AI.
A Few Common Questions I’ve Gotten:
1) Did you receive training?
Since I joined after the initial onboarding phase, I didn’t go through formal training. Most of the physicians had already been on the project for a few weeks and were welcoming and helpful to those of us just starting out.
2) Do you need a background in AI?
No. I don’t have one. A family member works in tech, so I’ve picked up some exposure through conversations, but no formal training. It’s more about applying clinical reasoning in a structured format.
3) There’s been debate in physician forums about whether tools like this might replace doctors someday. What are your thoughts?
Those implications are outside the scope of this piece. Personally, I believe technology will keep advancing whether we embrace it or not—so it’s in our best interest to engage early and help shape it for the better. I see these tools not as replacements but as support—helping us reduce administrative burdens, surface insights faster, and ultimately spend more time on patient care. My own view is that they’ll also improve access to care and raise the bar on quality.
