ML Research Engineer: Tuning & Inference Specialist

Mercor; United States; Remote; USD 180,000 - 240,000; Full time

Job summary

Mercor seeks an ML research engineer, research scientist or applied scientist to generate hard, long-horizon task ideas for a leading AI lab. You will prioritize inference, evaluation and systems work, with tasks split evenly between GPU and CPU execution, and explain why each problem challenges frontier models.

You primarily contribute ideas and do not need to build or review the tasks, focusing on clear explanations of model hardness and research direction impact.

Qualifications

• Current or recent role as ML research engineer, research scientist or applied scientist.
• Hands-on experience personally training, fine-tuning or optimizing models in PyTorch or JAX.
• Deep knowledge of LLMs, GPU inference and performance, RL, or scientific ML.
• Able to explain why a problem is hard for a frontier model.

Responsibilities

• Come up with ideas for hard machine learning engineering tasks for a long-horizon AI research project.
• Design tasks hard enough to stump frontier models; a strong task should be solved in few attempts.
• Lean toward inference, evaluation and systems work rather than heavy model training.
• Explain clearly why each problem is hard for a frontier model.
• You mainly contribute ideas and do not need to build or review the tasks.

Skills

ML research engineer

Research scientist

Applied scientist

PyTorch

JAX

LLMs knowledge

GPU inference & performance

Reinforcement learning

Scientific ML

Explain constraints for frontierModels

Education

PhD in ML, CS or physics

MSc + strong research output

Tools

PyTorch

JAX

Job description

Mercor seeks an ML research engineer, research scientist or applied scientist to generate hard, long-horizon task ideas for a leading AI lab. You will prioritize inference, evaluation and systems work, with tasks split evenly between GPU and CPU execution, and explain why each problem challenges frontier models.

You primarily contribute ideas and do not need to build or review the tasks, focusing on clear explanations of model hardness and research direction impact.