ML Benchmark Researcher
Job Description
Author and validate research-grade machine-learning benchmarks in areas where you write code and publish results.
What you will do
• Build realistic research tasks, reference solutions and grading rubrics.
• Validate whether benchmarks measure the intended capability and discriminate between correct and weak approaches.
• Contribute depth in areas such as language models, deep learning, RL, vision/generation, robotics, efficient systems, optimization/theory, classical or adaptive learning, time series, causal reasoning, trustworthy learning or AI for science.
Required background
• At least one year of hands-on ML research, with no upper experience limit.
• At least one first-author ML paper or paper in a closely related field.
• Completed or ongoing master’s/doctoral study in a quantitative subject such as machine learning, computer science, statistics or mathematics.
• Practical ML coding in training, evaluation or systems; annotation alone does not meet this research background.
Preferred background
• Multiple first-author papers or results demonstrating measured improvement over a baseline.
• Publications at venues including NeurIPS, ICML, ICLR, CVPR, ACL, EMNLP, CoRL or MLSys.
• Deep expertise in an applicable research area.
Location and eligibility
• Remote work; country restrictions are not stated.
Availability
• Flexible remote research work, approximately 5–40 hours per week.
Pay and engagement
$158/hr
USD hourly compensation. The applicable rate and agreed work scope are confirmed before work begins.