Training Turk

Excel Expert — General

$70–100/hr · Mercor · Hourly, 40 hours a week

You design, evaluate, and refine Excel tasks for AI training, leveraging deep cross-industry spreadsheet expertise.

What you would do

  • Guide research teams in closing knowledge gaps by designing realistic Excel problems drawn from your professional experience
  • Develop Excel tasks and write accurate, well-structured solutions across multiple business domains and functions
  • Evaluate AI-generated Excel solutions and tasks, providing clear written feedback on correctness and business applicability
  • Partner with other subject matter experts to maintain uniformity and quality across training datasets
  • Help AI systems improve by identifying patterns in where models struggle with formulas, data modeling, and spreadsheet logic

Who they want

  • Demonstrated advanced, professional-level Excel expertise with complex formulas, PivotTables, Power Query, Power Pivot, VBA, and dashboard design
  • Proven generalist experience applying Excel tools across many domains and operational functions, rather than single-function expertise
  • Available 40 hours per week through end of September, then 20 hours per week thereafter
  • Strong verbal and written communication skills, problem-solving ability, and interpersonal effectiveness
  • Teaching or training experience, or advanced degree (PhD), considered a significant plus; prior AI training or human data work also valued

Main skills

Advanced excel formulasDynamic arrays and array functionsPivottables and data modeling

What the interview asks about

  1. 1.Complex formula design across problem types

    Advanced formulas require knowing when INDEX/MATCH beats VLOOKUP, when dynamic arrays apply, and when a helper column is better.

    For example: “Design a task: given a table with employee names (duplicates), departments, and salaries, extract all salaries for a specific employee. What three approaches would you use?”

  2. 2.Cross-industry generalist perspective

    The role needs someone who built financial models, analyzed operations data, and designed dashboards; this breadth lets you recognize patterns.

    For example: “You worked in finance, supply chain, and marketing. Describe a spreadsheet problem you solved in one industry that taught you something reusable for the others.”

  3. 3.Identifying and explaining formula errors

    AI systems will generate incorrect formulas; you need to spot the error, explain why it is wrong, and articulate the correct approach.

    For example: “An AI generates =SUM($A$2:A2) for running totals. It works for rows 2-5 but fails on row 6. What is the bug and how would you explain the fix?”

  4. 4.Designing realistic business scenarios

    Training data quality depends on tasks that reflect actual work; you translate professional experience into clear, grounded problem statements.

    For example: “Design a task around inventory management. What real-world constraints would you build in like lead times or seasonal demand?”

  5. 5.Teaching Excel concepts to non-experts

    This role involves guiding research teams and explaining spreadsheet thinking to people who may not be Excel users; clarity is essential.

    For example: “A researcher doesn't understand Power Query. Explain why it is better than manual data manipulation for an ETL task.”

A task you may get

Design a multi-step Excel task for model training: a realistic business scenario that requires 2-3 connected formulas and basic data transformation. Write the scenario, the solution, and notes on what skills this tests.

How to prepare

  • Map your career: which industries did you work in, and what Excel techniques did each demand? Identify patterns that transfer across domains.
  • Study how teaching platforms explain Excel concepts; notice which explanations are clear and adopt those patterns in your feedback.
  • Review your own spreadsheets: take a file you built professionally and decompose it into training tasks for an AI model.
  • Research what errors AI systems commonly make with Excel: misapplied functions, off-by-one errors, incorrect relative/absolute references.

The facts

Pay
$70–100/hr
Hours
Hourly, 40 hours a week
Where
Remote · United States
Open to
USA
Field
Data Analysis
Posted
9/12/2026
Places left
10

We wrote this page from the public Mercor listing. It may be out of date, so read the full posting before you apply.