About Mercor's Hiring Philosophy

Mercor connects domain experts with AI labs that need high-quality human evaluation of model outputs. The General Evaluator Expert role sits at the center of that mission: you review, rate, and give structured feedback on AI-generated responses across a wide range of topics, flagging factual errors, reasoning gaps, safety issues, and instruction-following failures. Because the work feeds directly into model training and quality control, Mercor screens heavily for judgment, attention to detail, and clear written reasoning rather than for a specific degree or job title.

The hiring process itself is unusually lightweight and asynchronous. Nearly everyone applies online and is routed straight into an assessment, with little or no live human contact before an offer is generated. That speed is a double-edged sword: candidates praise how fast they can move from application to offer, but many report thin communication, unclear performance expectations, and difficulty getting questions answered. Go in prepared to be self-directed and to document everything.

Quick Stats

* Typical process: 1 to 2 stages (online application plus a timed assessment), roughly 1 to 3 weeks

* Format: Online application and remote asynchronous assessment (no traditional live panel for most candidates)

* Core focus: Evaluation judgment, attention to detail, written reasoning, instruction-following, rubric application

* Difficulty: Moderate (avg 3.0/5); the assessment is fair, but the unclear expectations and thin communication frustrate candidates

What Mercor Looks For

* Consistent, rubric-based judgment when rating AI outputs against clear criteria

* Sharp attention to detail and the ability to catch subtle factual or reasoning errors

* Clear, concise written explanations that justify each rating

* Reliability and self-direction in a remote, largely asynchronous workflow

"I did an assessment after I applied for a job and got an email to take the assessment. I satisfied every requirement for hiring and was formally offered the role." (General Evaluator Expert candidate, accepted offer)

Round 1: Online Application and Screening (~15 to 30 min)

What to Expect

The process begins entirely online. You apply through Mercor's platform, submit your background and any relevant expertise, and wait for an email routing you into the next stage. Every candidate in the data reported applying online, and the screening is automated rather than conversational. Expect to answer basic qualification and availability questions and to describe your relevant experience clearly, because there is often no recruiter call to clarify anything later.

Example or Reported Questions

* "What relevant experience do you have evaluating written or AI-generated content?"

* "What is your availability and expected weekly hours?"

* "Why do you want to work as an evaluator at Mercor?"

* "What subject areas or domains are you strongest in?"

Tips

* Be specific about your domains of expertise; the platform matches you to work based on what you list, so vague answers hurt you.

* Confirm logistics in writing early. Candidates report being ghosted, so keep a paper trail of your questions and their answers.

* Practice a crisp "why this role and why me" pitch with Nora's Standard Mode so your written and verbal responses stay tight and confident.

Round 2: Evaluation Assessment (~1 to 2 hours)

What to Expect

The core of the process is a timed, at-home assessment that simulates the actual job: you are given AI-generated responses (or prompts and outputs) and asked to rate them against a rubric, then justify your ratings in writing. This is where Mercor decides whether to extend an offer. One candidate completed it, met every requirement, and was formally offered the role directly off this stage. The assessment tests your accuracy, your consistency, and how clearly you can explain your reasoning under time pressure.

Example or Reported Questions

* "Rate this AI response against the provided rubric and explain your score."

* "Identify any factual errors or unsupported claims in the response below."

* "Which of these two responses better follows the user's instructions, and why?"

* "Explain what would make this answer higher quality."

Tips

* Read the rubric twice before you rate anything. Consistency against the stated criteria matters more than your personal opinion of quality.

* Keep written justifications short but concrete: name the exact error or gap rather than saying an answer "feels weak."

* Rehearse thinking out loud under time pressure with Nora's Technical Mode, treating each practice prompt like a live evaluation task so you can defend your scores quickly and clearly.

Round 3: Offer, KPIs, and Onboarding (~1 to 3 weeks)

What to Expect

If you pass the assessment, an offer is generated. This stage is administrative rather than interview-based, but candidates report that it is the most frustrating part of the process. Performance expectations and KPIs are reportedly not shared until after you accept, and communication can go quiet. One candidate who needed KPI details before accepting was ghosted, and the offer was withdrawn. Treat this stage as an active negotiation of clarity, not just a formality.

Example or Reported Questions

* "What are the performance metrics or KPIs I will be measured against?"

* "What is the pay structure and how are hours tracked?"

* "What does onboarding look like and when does work begin?"

* "Is there flexibility around workload or accommodations?"

Tips

* Ask for KPIs and success criteria in writing before accepting, and follow up politely but persistently if you are ghosted.

* If you need accommodations, request the specifics you need to make an informed decision; one candidate noted "Disabled people need to know if they'd be accepting a position at which they'd be at great risk of failure" (General Evaluator Expert candidate).

* Use Nora's Salary Negotiation Mode to rehearse asking about pay, hours, and expectations firmly without underselling yourself or stalling the offer.

Frequently Asked Questions (FAQ)

1) How many rounds are there?

Typically 1 to 2 real stages: an online application and screening, followed by a timed evaluation assessment. Offer and onboarding follow as an administrative step. There is usually no traditional live panel interview for this role.

2) What topics are most common?

* Rating AI outputs against a rubric and justifying scores in writing

* Catching factual errors, reasoning gaps, and instruction-following failures

3) How long does the process take?

Most candidates move quickly, often within 1 to 3 weeks from application to offer. Be aware that the post-offer stage can stall; one candidate was told a withdrawn offer would reopen in 3 weeks and it did not, so keep your own timeline and documentation.

4) How should I prepare?

* Practice evaluating written content against a rubric, focusing on consistency and concrete justifications rather than gut reactions.

* Sharpen your written reasoning so you can explain a score in one or two clear sentences under time pressure.

* Get every logistical and KPI question answered in writing before you accept, since expectations are reportedly shared late.

* Rehearse with Nora AI: use Standard Mode for the screening pitch, Technical Mode to drill timed evaluation tasks, and Salary Negotiation Mode to practice pinning down pay, hours, and KPIs before accepting.