Remote | Frontend Engineer (Web UI Evaluation) — Up to $80/hour

24-Mag • New York, NY

5d • Remote

About The Position

We are sharing a specialised part-time consulting opportunity for experienced frontend and full-stack engineers with strong HTML/CSS expertise, browser debugging skills, and an exceptional eye for visual and structural accuracy. This role supports advanced AI evaluation focused on frontend code generation and web-page replication. Selected engineers will compare AI-generated webpage implementations against reference designs, evaluate visual and technical fidelity, inspect the underlying code and responsive behaviour, and provide precise written feedback on which implementation performs better and why.

Requirements

• 3+ years of professional frontend or full-stack web development experience
• Strong proficiency with HTML and CSS
• Deep understanding of semantic markup, flexbox, CSS Grid, media queries, and responsive design
• Ability to read and reason about legacy float- and table-based layouts
• Strong working knowledge of browser developer tools, including element inspection, computed styles, the box model, and network inspection
• Sufficient JavaScript knowledge to understand scripts and DOM behaviour
• Comfortable using a terminal to run and inspect local web projects
• Strong ability to identify visual differences at a detailed, pixel-level standard
• Excellent written English and ability to justify technical evaluation decisions precisely
• Desktop or laptop capable of displaying a browser window at 1920px width or greater
• Administrator rights on the machine used for project work
• Ability to install and run a local development or static-server environment

Nice To Haves

• Experience translating Figma, PSD, or comparable designs into production frontend code is highly valued
• Prior code review, technical assessment, RLHF, preference evaluation, or AI model evaluation experience is advantageous
• Familiarity with WCAG, ARIA, accessibility testing, web scraping, or DOM parsing is beneficial
• Experience with common failure patterns in LLM-generated frontend code is highly valued

Responsibilities

• Compare a reference webpage against two AI-generated frontend implementations
• Render candidate pages side by side at desktop resolution and determine which more faithfully reproduces the target
• Evaluate overall visual similarity, layout accuracy, and implementation quality
• Identify significant defects and rank them by impact
• Flag ties, ambiguous comparisons, and technically broken evaluation items when appropriate
• Inspect hand-written HTML and CSS for structural and implementation accuracy
• Evaluate semantic markup, flexbox, grid, media queries, and traditional float- or table-based layouts
• Determine whether visual fidelity is achieved through robust layout techniques rather than fragile hardcoded positioning
• Identify excessive absolute positioning, unnecessary inline styles, and other implementation shortcuts
• Assess whether the underlying structure will remain correct outside a single viewport configuration
• Compare box-model behaviour, margins, padding, alignment, and spacing
• Evaluate typography including font family, size, weight, line height, and letter spacing
• Review colours, borders, backgrounds, images, and asset handling
• Inspect stacking order, overflow, clipping, and positioning behaviour
• Identify subtle visual discrepancies that affect replication quality
• Assess how candidate implementations behave when viewport dimensions change
• Evaluate whether responsive layouts use appropriate CSS techniques
• Review heading structure, landmarks, and semantic HTML quality
• Determine whether layouts remain stable and coherent across different dimensions
• Evaluate accessibility-related implementation quality where relevant
• Use browser developer tools to inspect DOM structure and computed styles
• Analyse the CSS box model and cascade to diagnose visual differences
• Use network and asset inspection where necessary to identify rendering issues
• Read JavaScript sufficiently to understand DOM manipulation and page behaviour
• Run provided webpages locally using terminal-based development workflows
• Write structured rationales explaining each preference decision
• Describe specific defects in clear, technically actionable language
• Distinguish major implementation failures from minor cosmetic differences
• Apply evaluation standards consistently across frontend tasks
• Contribute high-quality human preference data used to improve AI-generated frontend code

Benefits

• Part-time independent contractor engagement
• Fully remote
• Flexible scheduling based on project requirements
• Up to $80/hour compensation

What This Job Offers

• Job Type: Part-time
• Career Level: Mid Level
• Education Level: No Education Listed