Frontend Engineer — Web Replication Preference Rater

🏢 About Mercor

Mercor is an AI talent platform that partners with leading AI research labs and enterprises to train and evaluate frontier AI models using human domain expertise.

🎯 The Role

As a Frontend Engineer — Web Replication Preference Rater, you will build a high-quality human preference dataset on AI-generated frontend code. You will compare reference web pages against candidate AI replications, inspect underlying markup with browser DevTools, and write structured technical evaluations explaining where and why code falls short.

🛠️ Tech Stack

• Semantic HTML5 & Vanilla CSS
• Flexbox & CSS Grid
• Modern JavaScript (DOM manipulation)
• Browser DevTools (Computed Styles, DOM Tree, Box Model)
• Terminal & Local Server Tooling

💼 What You'll Do

• Render reference pages alongside candidate AI replications at desktop width (1920px+) and determine the closer reproduction.
• Audit layout fidelity in detail: box model, typography, spacing, colors, borders, z-index, and asset handling.
• Inspect computed DOM markup with DevTools to identify brittle code (hardcoded pixel offsets, absolute positioning, inline styles).
• Evaluate semantic quality, responsive layout choices, and fidelity to legacy float/table structures where applicable.
• Write structured technical rationales ranking identified defects in actionable, precise language.
• Flag ambiguous cases, broken test cases, or ties objectively rather than forcing arbitrary preferences.

✅ What You'll Need

• 3+ years of professional web development experience, with a focus on frontend or full-stack engineering.
• Deep fluency in hand-written HTML and CSS (semantic markup, Flexbox, CSS Grid, media queries, legacy layouts).
• Strong proficiency with browser DevTools (element inspector, computed styles, box model, network tab).
• Working knowledge of JavaScript to trace DOM mutations and script behavior.
• Comfort using the CLI to clone repositories and run local static servers.
• Excellent written English with the ability to write precise, actionable technical feedback.
• Laptop/desktop capable of rendering at 1920px+ width with admin permissions to run local servers.

🌟 Nice to Have

• Background in RLHF, preference labeling, or model evaluation datasets.
• Code review, technical writing, or design-to-code (Figma/PSD to HTML/CSS) experience.
• Web accessibility expertise (WCAG guidelines, ARIA attributes, screen reader behavior).
• Understanding of common LLM frontend code generation failure modes.