AI Trainer

Draft AI is hiring AI Trainers to work on paid engagements supporting the development and improvement of advanced AI systems.

AI Trainers work directly on projects designed to improve how AI models reason, follow instructions, complete complex workflows, and perform real-world professional tasks. Depending on the engagement, work may include evaluating model responses, identifying errors, ranking outputs, creating challenging prompts, completing domain-specific tasks, and providing detailed human feedback used in model post-training.

This is not a traditional software engineering role. We are looking for thoughtful, detail-oriented people who can apply strong judgment and clearly explain what makes an AI response correct, incorrect, or incomplete.

What you may work on:

• Evaluate and compare responses generated by advanced AI models
• Rank AI outputs based on accuracy, reasoning, clarity, and instruction following
• Create prompts and tasks designed to challenge model capabilities
• Identify factual, logical, and reasoning errors
• Complete real-world workflows that help train AI agents
• Provide structured feedback used for AI post-training and evaluation
• Work on specialized projects aligned with your academic or professional background

What we’re looking for:

• Strong analytical and written communication skills
• Excellent attention to detail
• Ability to follow detailed project guidelines consistently
• Comfort working with AI tools and learning new workflows quickly
• Strong judgment when evaluating ambiguous or complex responses
• Undergraduate degree in progress or completed, or equivalent professional experience
• Prior AI training or annotation experience is helpful but not required

Projects are remote and flexible. Engagement availability and hours vary based on client needs, qualifications, and performance. Selected applicants may be asked to complete a qualification or calibration assessment before joining a paid engagement.