Frontend Code Evaluation Specialist

Summary

About OpenTrain OpenTrain is the #1 platform for finding and building careers in AI training and data labeling. OpenTrain AI is recruiting contractors for specialized evaluation work that helps shape how advanced AI systems generate frontend code.Remote contractor opportunityWork with AI-generated frontend implementationsBuild experience in technical AI evaluationAbout AI Training and Preference Evaluation AI training is the human side of building artificial intelligence. People review examples, compare model outputs, and provide careful judgments so AI systems can become more accurate, useful, and reliable.In this role, your preference judgments help assess which AI-generated web implementation better reproduces a reference site. You will evaluate both what users see and how the code produces the result.Review and compare outputs from AI systemsProvide evidence-based human preference judgmentsContribute to the development of better coding modelsThe Role As a Frontend Code Evaluation Specialist, you will run a crawled reference website and two self-contained model replications locally. You will inspect screenshots and interactive states, then determine which candidate is the stronger reproduction.Your evaluation will cover rendered fidelity and implementation quality, including responsiveness, semantic structure, layout techniques, asset handling, and signs that a model used screenshot-based shortcuts instead of rebuilding the page.Pay: $90 USD per hourContractor and part-time engagementDefault commitment: about 40 hours per weekMinimum stated availability: 20+ hours per weekTypical evaluation unit: 2 to 3 hoursProfessional written English requiredWork is available only in eligible countries listed belowWhat You'll Do You will assess reference pages and candidate implementations at a precise viewport, interact with their controls, and inspect the source code behind each result. Your written judgments should explain the preference clearly and cite specific evidence from the page or implementation.Sound judgment is important when deciding whether a task is genuinely broken. The work requires attention to visual details, browser behavior, document structure, and the difference between a robust implementation and a superficial visual imitation.Render the reference page and both replications at 1920x1080Compare hover, click, and scroll states state by stateAssess spacing, typography, colors, borders, images, overflow, and z-orderEvaluate responsive behavior across the implementationsInspect HTML, CSS, and JavaScriptReview semantic structure and implementation qualityIdentify hardcoded offsets and inappropriate absolute positioningSpot undifferentiated markup and screenshots used instead of rebuilt sectionsWrite specific, evidence-cited rationales for each preferenceDetermine whether an evaluation task is genuinely brokenRequired Qualifications and Setup The role requires three to eight years of professional frontend or full-stack web development experience, although the opportunity is classified as entry level in the supplied role information. You should be comfortable evaluating both modern and legacy web structures and explaining technical conclusions in professional written English.You will need a desktop or laptop capable of displaying a 1920x1080 viewport, along with administrator rights to install and run a local server. The work also requires command-line and browser troubleshooting in a local development environment.Three to eight years of professional frontend or full-stack web development experienceStrong command of handwritten HTML and CSSKnowledge of semantic markup, flexbox, grid, media queries, float-based layouts, and table-based layoutsPractical browser DevTools experienceAbility to set exact viewports, inspect elements, and review computed stylesComfort with command-line workflows, archives, local static servers, and relative assetsAbility to troubleshoot broken image referencesEnough JavaScript knowledge to understand scripts and their effects on the DOMDesktop or laptop capable of displaying a 1920x1080 viewportAdministrator rights to install and run a local serverProfessional written English for detailed evaluation rationalesHelpful Background Previous experience with AI preference evaluation, RLHF, model evaluation, or structured code review can help you succeed. Experience in pixel-perfect design-to-code work and understanding how language models fail when generating code are also valuable.RLHF or preference labelingAI model evaluationStructured code reviewPixel-perfect design-to-code implementationAccessibilityWeb scraping and archivingDOM parsingUnderstanding common AI-generated code failuresThis opportunity is remote, but contractors must reside in an eligible country. Eligible locations include the United States, Canada, the United Kingdom, Ireland, France, Germany, Spain, Portugal, Italy, the Netherlands, Belgium, Austria, Denmark, Sweden, Finland, Norway, Poland, the Czech Republic, Slovakia, Hungary, Slovenia, Croatia, Romania, Bulgaria, Greece, Cyprus, Malta, Estonia, Latvia, Lithuania, India, Mexico, Guatemala, Costa Rica, Panama, the Dominican Republic, Colombia, Ecuador, Peru, Chile, Argentina, Uruguay, Paraguay, or Brazil.Applicants must reside in an eligible country#J-18808-Ljbffr

Job Description

About OpenTrain OpenTrain is the #1 platform for finding and building careers in AI training and data labeling. OpenTrain AI is recruiting contractors for specialized evaluation work that helps shape how advanced AI systems generate frontend code.Remote contractor opportunityWork with AI-generated frontend implementationsBuild experience in technical AI evaluationAbout AI Training and Preference Evaluation AI training is the human side of building artificial intelligence. People review examples, compare model outputs, and provide careful judgments so AI systems can become more accurate, useful, and reliable.In this role, your preference judgments help assess which AI-generated web implementation better reproduces a reference site. You will evaluate both what users see and how the code produces the result.Review and compare outputs from AI systemsProvide evidence-based human preference judgmentsContribute to the development of better coding modelsThe Role As a Frontend Code Evaluation Specialist, you will run a crawled reference website and two self-contained model replications locally. You will inspect screenshots and interactive states, then determine which candidate is the stronger reproduction.Your evaluation will cover rendered fidelity and implementation quality, including responsiveness, semantic structure, layout techniques, asset handling, and signs that a model used screenshot-based shortcuts instead of rebuilding the page.Pay: $90 USD per hourContractor and part-time engagementDefault commitment: about 40 hours per weekMinimum stated availability: 20+ hours per weekTypical evaluation unit: 2 to 3 hoursProfessional written English requiredWork is available only in eligible countries listed belowWhat You'll Do You will assess reference pages and candidate implementations at a precise viewport, interact with their controls, and inspect the source code behind each result. Your written judgments should explain the preference clearly and cite specific evidence from the page or implementation.Sound judgment is important when deciding whether a task is genuinely broken. The work requires attention to visual details, browser behavior, document structure, and the difference between a robust implementation and a superficial visual imitation.Render the reference page and both replications at 1920x1080Compare hover, click, and scroll states state by stateAssess spacing, typography, colors, borders, images, overflow, and z-orderEvaluate responsive behavior across the implementationsInspect HTML, CSS, and JavaScriptReview semantic structure and implementation qualityIdentify hardcoded offsets and inappropriate absolute positioningSpot undifferentiated markup and screenshots used instead of rebuilt sectionsWrite specific, evidence-cited rationales for each preferenceDetermine whether an evaluation task is genuinely brokenRequired Qualifications and Setup The role requires three to eight years of professional frontend or full-stack web development experience, although the opportunity is classified as entry level in the supplied role information. You should be comfortable evaluating both modern and legacy web structures and explaining technical conclusions in professional written English.You will need a desktop or laptop capable of displaying a 1920x1080 viewport, along with administrator rights to install and run a local server. The work also requires command-line and browser troubleshooting in a local development environment.Three to eight years of professional frontend or full-stack web development experienceStrong command of handwritten HTML and CSSKnowledge of semantic markup, flexbox, grid, media queries, float-based layouts, and table-based layoutsPractical browser DevTools experienceAbility to set exact viewports, inspect elements, and review computed stylesComfort with command-line workflows, archives, local static servers, and relative assetsAbility to troubleshoot broken image referencesEnough JavaScript knowledge to understand scripts and their effects on the DOMDesktop or laptop capable of displaying a 1920x1080 viewportAdministrator rights to install and run a local serverProfessional written English for detailed evaluation rationalesHelpful Background Previous experience with AI preference evaluation, RLHF, model evaluation, or structured code review can help you succeed. Experience in pixel-perfect design-to-code work and understanding how language models fail when generating code are also valuable.RLHF or preference labelingAI model evaluationStructured code reviewPixel-perfect design-to-code implementationAccessibilityWeb scraping and archivingDOM parsingUnderstanding common AI-generated code failuresThis opportunity is remote, but contractors must reside in an eligible country. Eligible locations include the United States, Canada, the United Kingdom, Ireland, France, Germany, Spain, Portugal, Italy, the Netherlands, Belgium, Austria, Denmark, Sweden, Finland, Norway, Poland, the Czech Republic, Slovakia, Hungary, Slovenia, Croatia, Romania, Bulgaria, Greece, Cyprus, Malta, Estonia, Latvia, Lithuania, India, Mexico, Guatemala, Costa Rica, Panama, the Dominican Republic, Colombia, Ecuador, Peru, Chile, Argentina, Uruguay, Paraguay, or Brazil.Applicants must reside in an eligible country#J-18808-Ljbffr