Artificial Intelligence Researcher | Remote

About the role

Contribute to research-grade physics benchmarks used to evaluate advanced language models. Solve, review, and validate complex physics problems across multiple specialized subfields. Work asynchronously with a fully remote, flexible schedule. Help produce fully human-verified reference solutions for frontier physics reasoning.

Responsibilities

• Solve, review, and validate complex physics problems across multiple specialized subfields.
• Produce fully human-verified reference solutions for frontier physics reasoning.

Requirements

• Strong expertise in adversarial training of image classifiers and evaluating robust accuracy.
• Experience in training image classifiers for fine-grained recognition and model compression techniques.
• Hands-on experience with training image generative models and iterating against sample-quality metrics.
• Experience with supervised fine-tuning and preference optimization of language models.
• Experience in training multilingual or low-resource-language models and tokenizer design.

Independent contractor engagement with opportunities for project extensions.