AI Quality Assurance Jobs

## AI Image Reviewer / Quality Control Freelancer I am looking for a reliable and detail-oriented freelancer to review batches of AI-generated images and perform quality control. The job consists of reviewing the images **one by one** and deciding which images should be kept and which should be rejected. ### What you will check Images should be rejected when they contain noticeable generation problems, including: - Anatomical errors - Malformed hands, fingers, arms, legs, feet, etc. - Missing or duplicated limbs - Incorrect or distorted facial features - Strange body proportions - AI artifacts or visual glitches - Incorrect character features - Inconsistent clothing or accessories - Unwanted objects or elements - Major composition problems - Any other obvious issue that makes the imag...

Senior Software Engineer – AI Model Output Evaluation & Labeling We need a Senior Software Engineer to support our client's AI model labeling work. You'll evaluate AI model outputs and write clear rationales for each evaluation. You will work remotely on the client's computer via VM. Requirements: - Senior Software Engineer (5+ years preferred) - Native-level English writing skills (critical — rationales must be precise and grammatically flawless) - Comfortable working in a remote VM environment - Strong attention to detail and ability to follow guidelines - Preferred: Based in the Philippines - Prior AI/ML labeling or model evaluation experience is a plus Details: - Rate: $10/hr - Type: Ongoing contract - Location: Remote (VM access provided) - Start: Im...

Content Production & Product Operations Specialist - Canva, ElevenLabs, Web App QA We are preparing the first program for a new guided learning/self-development web application aimed initially at working professionals. I am looking for one reliable freelancer who can own the operational production workflow, rather than separate freelancers for design, audio, content upload and testing. The educational concepts, written content, scripts, scenarios and storyboards will primarily be provided. Your responsibility will be turning those specifications into polished production assets and configuring/testing them inside the web application. Main responsibilities Media production * Create polished educational/knowledge videos in Canva * Establish and maintain a consistent media style/temp...

Build WhatsApp-First B2B RFQ Marketplace with AI Supplier Matching PROJECT I need an experienced developer/team to build and launch a WhatsApp-first B2B RFQ and supplier-matching platform initially for Oman. I already have an Excel database of approximately 20,000 businesses containing company names, phones, emails, products, services, categories and locations. Every company can be BOTH buyer and supplier. Example: Buyer WhatsApps: “Need 3 forklifts in Sohar for 6 months.” System: WhatsApp → AI understands request → creates RFQ → searches database → ranks suppliers → sends RFQ → suppliers respond → buyer receives responses. Goal: maximum automation and minimal manual operation. --- 1. WHATSAPP BUSINESS Use official Meta WhatsApp B...

Indoor Moderated Video Data Collection Study - SAN FRANCISCO - $150 Project Overview: [Face Image/Mask Creation] We are seeking participants for a large-scale, moderated, in-person video data collection study designed to support the development and evaluation of AI technologies. This is a paid indoor and outdoor study where participants will record short selfie videos while performing simple, natural actions such as sitting, standing, lying down, and moderate/brisk walking. All recordings will be completed using secure, client-provided devices under the guidance of trained moderators. As part of this study, participants will also be asked to have their photos taken. These photographs will be used to create physical and digital facial replicas, including 3D-printed, latex, silicone, pap...

we were looking for someone to help us compare AI generated images, videos n voice clips and say which one is better, and why. thats the whole job honestly, but the "why" is the part that matters. you get a prompt, sometimes a reference photo, and two outputs. You go through the prompt like a checklist, check both against it, look at quality n artifacts, pick A or B, and write 2 to 3 sentences pointing at the exact thing that decided it. "B misses the third seedling" not "B felt weaker". For voice tasks you actually listen, not just read the transcript. what we need: native english speaker, this ones a must. Some background in AI training data, labeling / annotation or AI evaluation work. But honestly the mindset matters more than the years... you pick thing...

I have a quick, one-off assignment for U.S.-based freelancers who already have access to—or can easily sign up for—Anthropic’s Claude AI. Your role is simply to run through a short set of prompts I will supply, capture the responses, and share a brief reflection on how the tool behaved. No coding, no system integration—just hands-on interaction that helps me gauge everyday usability. Deliverables • The raw output Claude returns for each prompt (copy-paste or screenshots are fine) • A concise paragraph noting any quirks, strengths, or surprises you observed The entire task typically takes 20–30 minutes. A stable internet connection and a U.S. IP address are the only technical requirements. Please mention “ClaudeUSA” somewhere in your b...

We want to develop an *AI-powered Verification Copilot for GSCS INTERNATIONAL* to assist qualified verifiers throughout the complete SLCP/FSLM verification process on Worldly. ### Core Concept This is *not a general chatbot and not an autonomous verifier*. > *AI prepares. AI checks. AI alerts. Verifier decides.* ### How It Will Work The verifier selects the *country and verification/assessment period. The Copilot then supports and checks the **entire verification across all applicable FSLM questions*, not only a fixed number of questions. For each question and across the verification as a whole, the AI will: * Suggest appropriate responses based on GSCS methodology and available evidence * Apply applicable country laws and regulations * Check mandatory and unanswered questions * ...