Remote Physicist for AI Model Evaluation & Problem Solving

SaidGig; New York (NY); Remote; USD 83,000 - 165,000; Full time

Benefits offered by this job

Fully remote contract

Independent contractor arrangement

Job summary

SaidGig is seeking a highly capable physicist to help advance large language models in a fully remote contractor role. The work involves designing and solving advanced physics problems, producing clear step-by-step solutions, and providing detailed annotations to guide model evaluation.

You'll collaborate with AI researchers to align problem sets with evaluation goals, develop physics benchmarks from undergraduate through PhD levels, and translate complex ideas into accessible explanations with

Qualifications

• Pursuing master's/PhD/postdoc in Physics or related field.
• Ability to analyze and solve complex physics problems.
• Excellent written communication and collaboration for remote work.

Responsibilities

• Design and solve challenging physics problems that test LLMs.
• Produce accurate, high-quality step-by-step solutions with reasoning.
• Collaborate with researchers to align problem sets with evaluation goals.
• Help develop physics benchmarks across undergrad to PhD curricula.
• Translate complex physics concepts into clear explanations with visuals.
• Provide constructive feedback and detailed annotations.

Skills

Strong research

Analytical skills

Creative thinking

Lateral-thinking

Excellent written communication

Remote collaboration

Education

Master's degree or PhD in Physics or related field

Job description

SaidGig is seeking a highly capable physicist to help advance large language models in a fully remote contractor role. The work involves designing and solving advanced physics problems, producing clear step-by-step solutions, and providing detailed annotations to guide model evaluation.

You'll collaborate with AI researchers to align problem sets with evaluation goals, develop physics benchmarks from undergraduate through PhD levels, and translate complex ideas into accessible explanations with