How Our Screening Process Works — And Why It's Designed This Way

We get a lot of questions about the screening process. The most common one is: "What exactly is this going to be like?" It is a fair question. Most professional screening processes are opaque by design, which creates anxiety and, often, people dropping out before they even try. We want to be different, so here is an honest, detailed account of every step.

Step 1: Apply

The application is intentionally minimal. We ask for your name, email address, the discipline you are applying in, and a brief description of your professional background. No CV required, no portfolio, no cover letter.

Why so minimal? Because the application is not where we evaluate expertise. Resumes are unreliable signals of the specific competencies we need (ability to evaluate AI output, communication clarity, domain depth). The screening process is where we assess those things, and doing it there is fairer to everyone.

Within a few minutes of applying, you will receive a confirmation email with a link to the AI interview. The clock on the interview link is 7 days, so you can take it at a time that works for you.

Step 2: The AI Interview

The interview is voice-based, runs 20 to 25 minutes, and can be taken from any device with a microphone. It is led by an AI system, not a human interviewer.

This is not a quiz about AI. We are not testing whether you know what a transformer is or what RLHF stands for. The interview is a structured conversation about your domain. A medical professional will be asked about clinical topics. A lawyer will be asked about legal concepts and practice. An engineer will be asked about technical problems.

What the system is evaluating is three things: the depth of your knowledge (can you engage with nuance, or only with surface-level topics?), your communication clarity (can you explain complex ideas to someone who is not an expert in your field?), and your reasoning style (when you reach a judgment, can you explain why?).

Common questions include: "Walk me through how you would approach [domain-specific scenario]." "Explain [complex concept] to someone with a strong general background but no expertise in your field." "Describe a situation from your professional experience where you had to make a judgment call with incomplete information."

Step 3: The Case Study

After your interview, you will receive a written case study. It is a 30 to 45 minute exercise, and you have 72 hours from receipt to complete it.

The case study presents a realistic scenario from your discipline. A legal professional might receive a draft contract clause and be asked to identify potential issues. A medical professional might receive a clinical vignette and be asked how they would approach management. A software engineer might receive a code snippet and be asked to identify problems.

There is no single correct answer. What reviewers are looking for is the quality of your reasoning, the precision of your language, and the completeness of your analysis. An answer that identifies 8 of 10 relevant issues and explains each of them clearly is better than an answer that names all 10 but provides only superficial justification for each.

Step 4: Matching

Within 48 hours of submitting your case study, you will receive your result by email. Candidates who pass are immediately matched to active roles that align with their background.

The match is not random. It considers the specific domain you applied in, the subspecialty depth demonstrated in your interview and case study, and the tasks currently available. A contracts attorney is matched to contract review tasks. An oncologist is matched to oncology evaluation tasks.

If you do not pass, you will receive a specific explanation. The most common reasons are: insufficient depth in the case study response, communication clarity issues in the voice interview (not domain knowledge), and technical issues with the interview recording. You can reapply after 30 days.

Common Concerns

"I am not an AI expert." You do not need to be. The entire point of this work is that your domain expertise is what matters, and you have that. The AI context will be explained as part of your role orientation after matching.

"Is this legitimate?" The work produces training data for commercial AI applications. The end clients are large technology companies building AI products in your domain. Your work is covered by a standard NDA. Nothing you produce is attributed to you publicly.

"What if I fail?" You can reapply in 30 days. Many of the contributors currently active on the platform did not pass on their first attempt. The screening is genuinely evaluative, and we do not pass everyone, but it is not designed to exclude capable experts.