Applied Scientist, Robotics & Model Intelligence

Generalist; United States; Hybrid; USD 180,000 - 240,000; Full time

Job summary

Generalist is seeking an Applied Scientist to advance dexterous manipulation capabilities across diverse applications. You will shape evaluation, run experiments, and drive post-training improvements that generalize across industries.

The role blends research, engineering, and operations, demanding rigorous experimentation and clear communication with teams and partners. You will design robust task suites, manage data collection and model fine-tuning at scale, and push the boundaries of embodied

Qualifications

• PhD or equivalent research experience in machine learning, robotics, or a closely related field.
• Experience designing post-training recipes for dexterous manipulation tasks.
• Strong background in evaluating ML/robotics systems at scale and reporting results clearly.
• Excellent communication with research teams and partner-facing stakeholders.

Responsibilities

• Own end-to-end improvement of target capabilities, from baseline to mastery.
• Design task suites and evaluation protocols that generalize beyond a single deployment.
• Coordinate data collection, training and fine-tuning, and orchestrate large-scale evaluations.
• Run rigorous experimental analyses across model variants and data configurations to diagnose failures.

Skills

Data analysis at scale

Evaluation & benchmarking

Hands-on with robots

Machine learning / robot learning

Communication of findings

Education

PhD in ML/robotics or related field

Tools

Experimentation frameworks

Job description

Generalist is seeking an Applied Scientist to advance dexterous manipulation capabilities across diverse applications. You will shape evaluation, run experiments, and drive post-training improvements that generalize across industries.

The role blends research, engineering, and operations, demanding rigorous experimentation and clear communication with teams and partners. You will design robust task suites, manage data collection and model fine-tuning at scale, and push the boundaries of embodied