Retail Specialist

What you would do

• Teach research and engineering teams how retail merchandising, category management, and operations decisions are made in practice
• Write realistic, scenario-based retail problems with detailed, well-reasoned solutions rooted in actual merchandising logic
• Evaluate AI model outputs by comparing them against rubrics that measure correctness, judgment quality, and reasoning clarity
• Develop and continuously refine scoring rubrics tailored to specific retail tasks and decision contexts
• Collaborate with fellow domain experts to ensure consistency and rigor across all training data

Who they want

• Minimum 8+ years dedicated retail experience at recognized tier-one organizations (Amazon, Walmart, Target, Nike, Costco, Home Depot level)
• Demonstrated career progression from hands-on roles through management or specialism (e.g., Category Manager to Director)
• Prior hands-on experience evaluating AI or LLM outputs against structured scoring criteria; this is mandatory
• Strong verbal and written communication ability to explain retail reasoning and provide actionable feedback
• Reliable availability for 35+ hours per week during weekday business hours

Main skills

What the interview asks about

1. 1.Real-world retail problem authoring

   AI training requires problems grounded in actual scenarios that SMEs would solve daily; contrived or oversimplified tasks waste training data

   For example: “You notice a category's margins have compressed 4 percentage points while units sold rose 15 percent. Outline the analytical steps a Category Manager at a major retailer would take, and what actions might follow.”

2. 2.AI output evaluation against rubrics

   Scoring consistency and fairness is critical; vague or inconsistent rubrics undermine model training and introduce bias

   For example: “An AI system recommends shifting 30 percent of a seasonal fashion category's allocation to an adjacent size run based on two weeks of sell-through data. How would you structure a rubric to score this recommendation's soundness?”

3. 3.Merchandising judgment articulation

   Engineers cannot extract nuanced decision logic from retail experts unless that logic is articulated clearly and completely

   For example: “Describe a pricing decision you once made that balanced margin protection, competitive dynamics, and inventory turnover in a mature category. What trade-offs mattered most and why?”

4. 4.Cross-domain knowledge transfer

   Success requires translating retail complexity into explanations engineers and researchers can learn from and encode into AI systems

   For example: “A researcher asks why a retailer would reduce order quantities to a supplier even though forecasts show rising demand. Explain the inventory, cash flow, and risk factors at play.”

5. 5.Operational consistency in training data

   Inconsistent solutions or feedback teach AI systems noise instead of patterns; SMEs must catch and flag their own or peers' deviations

   For example: “You spot two similar markdowns coded differently by another SME - one flagged as a clearance decision, the other as a competitive response. How would you approach alignment?”

A task you may get

Develop a complex retail scenario involving category repricing under competitive pressure, margin erosion, and inventory constraints; write a detailed solution and rubric for scoring model responses.

How to prepare

• Review and document your most complex merchandising or category decisions in your career, capturing the reasoning and outcomes
• Study the structure of rubrics used in academic testing, certifications, or prior AI evaluation work
• Prepare examples of how you would explain retail decision-making to someone without retail background
• Reflect on instances where AI recommendations or algorithmic decisions missed important retail context or judgment calls

The facts

Pay

$60–80/hr

Hours

Part time, 35 hours a week

Where

Remote · United States

Open to

USA

Field

Business Operations

Posted

7/10/2026

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.