Saved to your account, encrypted. The AI that reads it doesn't keep it or train on it.
Benefits
Flexible schedule
Flexible working hours; choose preferred working days and hours, including weekends.
Other
Fully remote; global location.
Full job description
From Weekday AI’s careers site
This role is for one of our clients
Compensation: $60–$120 per hour Location: Global – Fully Remote Employment Type: Independent Contractor Work Commitment: Approximately 15 hours per week Schedule: Flexible – Choose your preferred working days and hours, including weekends.
About the Role
We are looking for highly skilled Data Scientists, Machine Learning Practitioners, and Applied Mathematics Experts with a strong track record in competitive data science, particularly on Kaggle or similar competition platforms.
Show full descriptionShow less
In this role, you will leverage your competition experience, mathematical reasoning, and programming expertise to develop challenging technical tasks for AI training and evaluation. The focus is on creating problems that require genuine analytical thinking and robust implementation rather than solutions that simply appear plausible.
Each task should include a reproducible execution environment, a verified reference solution, and reliable evaluation mechanisms to determine whether a solution is technically correct and complete.
The ideal candidate combines strong competitive programming and data science skills with the ability to develop, test, debug, and independently validate complex solutions.
Requirements
Key Responsibilities
Build reproducible computational environments, prepare relevant datasets and dependencies, and implement reliable reference solutions.
Design challenging data science and machine learning tasks that require strong analytical reasoning and practical implementation skills.
Develop test cases, evaluation criteria, and verification mechanisms that distinguish correct solutions from incomplete, inaccurate, or flawed approaches.
Execute coding agents against assigned tasks, analyse their outputs, investigate failures, and refine tasks based on observed performance.
Identify implementation errors, incorrect assumptions, data leakage, methodological weaknesses, and other issues that may affect solution accuracy.
Validate solutions through independent testing, computational experiments, and reproducible workflows.