If you care about clean markup, accessible interfaces, and frontend code that doesn’t fall apart in production, this role puts your standards to work. You’ll pressure-test AI models with real web development scenarios, verify technical accuracy, and document where the model’s reasoning breaks so the system gets better.
About Invisible Agency
Invisible supports companies by building scalable, high-quality operations. This freelance AI Trainer project uses expert-led evaluations to improve AI performance with rigorous, real-world training data.
Schedule
Remote (United States).
Contract / freelance.
Hours vary based on your availability (you’ll report average weekly hours).
You provide your own secure computer and high-speed internet. No company benefits (contractor role).
What You’ll Do
Challenge AI models with software engineering tasks and technical scenarios using HTML (and frontend web development workflows)
Assess outputs for correctness, clarity, maintainability, and standards compliance (semantic HTML, accessibility, performance)
Evaluate related JavaScript/TypeScript logic when solutions require interactivity or integrations
Capture reproducible error traces and failure modes, then recommend improvements to prompts and evaluation metrics
Document findings clearly, showing your work and reasoning so results are actionable
What You Need
Strong HTML proficiency, including semantic structure and accessibility best practices
Professional experience with JavaScript and/or TypeScript (required beyond basic markup work)
Ability to evaluate real-world web engineering concerns (async behavior, API integration, secure patterns, debugging)
Clear written communication and the ability to explain decisions, tradeoffs, and fixes
Bachelor’s/master’s/PhD in CS/SE (nice), or equivalent hands-on experience
Bonus signals: technical writing, open-source contributions, production frontend experience
Benefits
Pay range: $6–$65/hour depending on experience, expertise, and location
Remote, flexible contract work
High-impact work: you’re improving AI’s reliability for real engineers, not demo code
Quick reality check: this isn’t “I can write HTML.” It’s “I can judge whether a solution is correct, safe, accessible, and production-ready.” If that’s you, apply and set your rate accordingly.
Happy Hunting,
~Two Chicks…