frontend developer for AI-generated web interfaces

The opportunity focuses on evaluating AI-generated web interfaces and code by comparing frontend implementations with real-world reference websites and assessing visual fidelity and underlying code quality.

• Render a reference webpage and two AI-generated implementations at a 1920×1080 viewport and evaluate which version most accurately reproduces the reference;
• Compare visual fidelity across layout, spacing, typography, colors, borders, images, layering, and overflow behavior;
• Review source code and assess HTML structure, CSS architecture, responsiveness, maintainability, and implementation quality;
• Identify technical shortcuts such as excessive hardcoded positioning, unnecessary absolute positioning, poorly structured markup, or image-based recreations;
• Test responsive behavior across different viewport sizes and identify layout or usability defects;
• Use browser DevTools to inspect elements, computed styles, DOM structure, and interactive states;
• Provide specific, evidence-based written justifications explaining why one implementation is stronger than another;
• Identify genuinely broken evaluation tasks while distinguishing them from tasks that are simply challenging.

• 3–8 Years of professional web development experience, primarily in frontend or full-stack engineering;
• Strong hands-on experience with HTML, CSS, JavaScript, responsive web development, and modern frontend engineering practices;
• Deep understanding of semantic HTML, Flexbox, CSS Grid, media queries, and responsive layouts;
• Ability to read and understand both modern and legacy web layouts, including float-based and table-based designs;
• Strong proficiency with browser Developer Tools, including DOM inspection, computed styles, viewport testing, and interactive state debugging;
• Comfortable working from the command line, including extracting archives, running local development servers, and troubleshooting asset paths;
• Sufficient JavaScript knowledge to understand how scripts interact with the DOM and user interactions;
• Strong professional written English and the ability to provide clear, detailed technical justifications;
• Strong attention to detail and the ability to identify subtle visual and structural differences;
• Desktop or laptop capable of displaying a 1920×1080 viewport;
• Ability to run a local development environment and static web server;
• Administrator access to your own computer for installing required development tools;
• Practicing web developer with strong frontend and UI implementation expertise;
• Backend-only, ML/data-science-only, mobile-native-only, and DevOps-only backgrounds are generally not suitable for this evaluation work;
• Nice to have: Prior experience with RLHF, preference labeling, AI model evaluation, structured code review, or human-feedback projects, pixel-perfect design-to-code experience, web accessibility, familiarity with common failure patterns in AI-generated frontend code, web scraping, website archiving, DOM parsing, complex crawled website structures, previous AI evaluation projects, multi-hour evaluation sessions.

• Freelance hourly contract;
• Part-time employment;
• Each evaluation unit typically takes 2–3 hours and is completed within a defined time window;
• Requires the ability to dedicate uninterrupted blocks of time to individual evaluations;
• A functioning local development environment is required;
• Competitive compensation of $90/hour;
• Independent contractor engagement;
• Compensation is based on completed evaluation work in accordance with project requirements.