Mathematics Expert - AI Training
Braintrust · India
- Experience5–8 yrs
- SalaryDisclosed
- Work moderemote
- Posted1 Oct 2026
About Braintrust
Braintrust is hiring in India in technology software. This role looks for around 5+ years of experience.
Skills
- mathematics
- analytical reasoning
- AI evaluation
- data annotation
- English
The role
A mathematics expert at a technology marketplace develops mathematical reasoning for AI training, evaluates large language models, and provides structured feedback. The role applies AI evaluation and data annotation to assess symbolic manipulation, abstraction, and multi-step reasoning.
Full job description
Use advanced mathematics and analytical reasoning to help improve large language models. You will probe model limitations and explain complex concepts through clear, step-by-step reasoning.
What You Will Do
Create high-quality, step-by-step solutions with detailed reasoning. Work with AI researchers to align task designs with evaluation goals, including symbolic manipulation, abstraction and multi-step reasoning. Construct novel scenarios and evaluate complex reasoning pathways. Provide structured feedback and detailed annotations to improve model quality.
What You Will Bring
A PhD in Mathematics or an equivalent technical field. Strong mathematics knowledge spanning advanced engineering-entrance through graduate and PhD-level concepts. Strong analytical, research and problem-solving skills with excellent English comprehension. Exceptional structured written communication and the ability to provide constructive feedback remotely. Lateral thinking and the ability to work independently in a fast-paced remote setting. A personal desktop or laptop and stable high-speed internet.
Experience in AI evaluation, data annotation, content review, quality assurance or a related analytical role is preferred, not required.
PROJECT DETAILS
Expected engagement length is 12 weeks, with an immediate start. A minimum of 20 hours per week, up to 40 hours per week, is required. Four hours of daily Pacific Time overlap is required. Selection includes an assessment and a delivery review.
PAYMENT
Turing pays $80 USD directly for each approved task.