Jobs at Careerflow.ai

Remote annotator reviews Japanese text and images, draws bounding boxes, answers structured questions, and writes short summaries per guidelines to train AI systems. On-the-job training provided; assessment required.

Evaluate AI-generated responses for business users using your corporate Gmail history. Assess relevance, personalization, accuracy, and helpfulness; compare responses, identify personalization errors, provide detailed feedback, follow guidelines, and protect data privacy while working remotely.

Remote Indonesian-language annotator role: review and label text and images, draw bounding boxes, answer structured questions, write short summaries, and follow annotation guidelines to train AI. Long-term (minimum 6 months), 30-40 hours/week, paid hourly; on-the-job training provided and a short language assessment is required.

Evaluate AI-generated responses for business users by comparing outputs against professional email history, assessing relevance, accuracy, personalization, and providing detailed feedback to improve models while maintaining data privacy.

Freelance remote role reviewing images in an online tool to rate technical quality (blur, noise, exposure, compression). Apply guidelines consistently, submit accurate, detail-oriented evaluations, and commit ~20+ hours/week for several weeks.

Review AI training problems and research environments to assess realism, design quality, and grading accuracy (do not solve). Provide constructive feedback and detailed annotations for 30–90 minute tasks in a remote, contractor capacity.

Evaluate AI-generated responses using your personal Gmail account to assess relevance, personalization, accuracy, and helpfulness across English and German. Provide detailed feedback, compare responses, identify incorrect personalization or unsupported assumptions, and follow confidentiality and consent procedures.

Evaluate AI-generated responses for business users using your corporate Gmail history. Assess relevance, accuracy, personalization, and identify incorrect personalization or unsupported assumptions. Compare responses, provide clear, structured feedback, follow evaluation guidelines, and maintain data privacy. Work remotely full-time (4–40 hours/week) with required overlap with PST for a 16-week contract.

Evaluate Gemini personalization by designing multi-turn prompts based on personal context, scoring responses for relevance, personalization, hallucinations and grounding, validating data usage, ranking outputs, writing rationales, and maintaining data hygiene while collaborating with remote teams.

Build pixel-perfect, responsive web components from style guides and screenshots using TypeScript and CSS3. Ensure accessibility and design-system conformity, optimize UI performance, document work asynchronously, resolve design ambiguities, and contribute domain expertise to train AI systems.

Evaluate AI-generated responses for business users by reviewing personalization against your professional Gmail history. Rate relevance, accuracy, helpfulness, and personalization; compare responses; identify incorrect personalization or unsupported assumptions; provide clear, structured feedback while following guidelines and maintaining data privacy.

Remote Danish language annotator to review and label text and images, draw bounding boxes, answer structured questions, and write short summaries following annotation guidelines. No prior experience required; on-the-job training provided. 30–40 hours/week, $8/hour, minimum 6-month commitment.

Produce 1-5 minute walk-and-talk videos showcasing public locations. Speak naturally on camera, record steady footage with clear audio, follow creative guidelines, submit location details, respect privacy and safety, and deliver videos on schedule as a freelance contractor.

Review and annotate Chinese text and images to train AI: draw bounding boxes, answer structured questions, write short summaries, and follow annotation guidelines. Remote, long-term contract (minimum 6 months), 30–40 hours/week, paid hourly with a short language assessment required.

Create 1–5 minute walk-and-talk videos of public city locations, speak naturally on camera, record steady footage with clear audio, submit videos with accurate location details, follow creative guidelines, respect privacy and safety, and meet weekly time commitments.

Design and build reinforcement-learning training environments and diverse tasks to evaluate and improve LLM agents; iterate rapidly on task designs from customer feedback, deliver high-quality outputs with minimal supervision, and maintain PST overlap for collaboration.

Develop and maintain MCP tools and the MCP Toolbox for Databases; expand test coverage for agentic coding assistants; refactor and maintain test frameworks; convert technical posts into production-ready sample code and documentation; build real-world data applications (QueryData) to support product R&D and collaborate with cross-functional teams to accelerate agentic AI adoption.