$100–150/hr
The role in one line
You evaluate AI-generated technical documents, slides, and spreadsheets for accuracy, usability, and quality in software engineering contexts.
Written by Training Turk from the public listing; it may be incomplete or out of date. Read the full posting on Mercor.
What you would do
- Review AI outputs including presentations, data sheets, and technical documentation for correctness and clarity
- Identify factual errors, aesthetic problems, visualization issues, and structural gaps in generated artifacts
- Assess whether materials would be usable by actual engineers and meet professional standards
- Provide detailed, structured written feedback explaining specific problems and improvement recommendations
- Submit evaluations with supporting evidence and technical reasoning
Who they are looking for
- 5+ years professional software engineering experience at top-tier firms in US, UK, Canada, Australia, or New Zealand
- Native or professional fluency in English
- High proficiency in Microsoft Office including PowerPoint, Excel, and Word
- Proficiency with Google Workspace tools for documents and spreadsheets
- Advanced degree preferred from reputable institution
What the interview is likely to probe
1.Subtle technical inaccuracy detection
An AI system might correctly describe core concepts but introduce subtle errors in edge cases or implementation details that a practicing engineer would catch.
Expect something like: “An AI-generated architecture diagram shows a microservices deployment with inter-service communication but omits failure scenarios and retry logic. How critical is this omission and what feedback would you give?”
2.Aesthetic and presentation judgment
Professional technical communication requires correct chart types, readable layouts, and proper data representation; AI often gets this wrong in ways that hurt usability.
Expect something like: “A spreadsheet compares 8 performance metrics across 5 systems using the default chart type. What problems do you spot and what would you recommend?”
3.Audience and context alignment
An artifact might be factually correct but inappropriate for its audience or use case; your feedback must reflect the actual context.
Expect something like: “An AI generated a deep technical specification that's accurate but contains implementation details a product manager presenting to stakeholders doesn't need. How do you evaluate this?”
4.Structural and logical soundness
Beyond surface errors, documents and slides need sound logical flow, correct sequencing, and clear narrative; AI often skips these higher-level qualities.
Expect something like: “A slide deck on system design covers all necessary topics but presents them in a non-linear order that makes the argument hard to follow. Walk through your feedback approach.”
5.Comparative evaluation across artifacts
Your feedback quality depends on understanding what good looks like; you need to distinguish between an artifact that's merely acceptable versus one that's excellent.
Expect something like: “You're evaluating two AI-generated architecture diagrams covering the same system. How do you determine which is better and what specific feedback would you provide to each?”
Exercise you may get
Evaluate a 5-slide AI-generated deck on a technical topic, identify 3-5 errors or quality issues (factual, aesthetic, or structural), and provide detailed feedback with improvement suggestions.
How to prepare
- Collect examples of well-designed technical documentation and slides from your experience
- Identify 3-5 common errors you see in presentations or technical writing
- Review Microsoft Office and Google Workspace features for creating professional technical materials
- Prepare specific examples of how technical errors change usability or credibility
Facts
- Pay
- $100–150/hr
- Commitment
- hourly
- Hours
- 40 per week
- Work arrangement
- remote · Remote
- Domain
- Software Engineering
- Company
- Dorado
- Posted
- 9/21/2026
- Open slots
- 51