About the role
This is a 1099 independent contractor opportunity with Scale AI's Human Frontier Collective, not traditional employment. Accepted Fellows are matched on a rolling basis to AI-training projects where they evaluate model outputs, review machine-learning code, and help build the data and guardrails used to train frontier AI systems. The engagement is fully remote within the US, self-scheduled, and continues only as long as there's mutual interest on both sides.
Responsibilities
- Get matched to partner projects building AI safety and policy guardrails
- Design, review, and optimize PyTorch models for client and partner teams
- Evaluate machine-learning code and AI-generated implementations for correctness and efficiency
- Evaluate AI model outputs and help design evaluation experiments
- Build reinforcement-learning (RL) environments used in model training
- Co-author technical reports and research papers summarizing findings
- Set your own weekly schedule — Fellows typically work 10–25 hours per week
Requirements
- This is a 1099 independent contractor engagement, not an employee role, with no fixed end date
- PhD or postdoctoral experience in machine learning or computer science, or senior industry ML experience
- Strong Python skills with hands-on experience in modern ML frameworks such as PyTorch or TensorFlow
- AWS, Docker, or LangChain experience is a plus
- Must hold, or be able to independently obtain, full-time US work authorization — Scale does not sponsor visas for this engagement
- Fully remote with no minimum weekly-hour commitment beyond mutual interest in continuing
- Detail-oriented and research-minded, comfortable with applied AI/ML evaluation work
- Pay is project-dependent and disclosed only after matching to a project, not a fixed hourly rate
What we offer
- Access to an expert researcher network, newsletters, and discussion channels
- Invitations to virtual sessions and in-person community events
- Opportunity to co-author published research papers