2027 Summer Intern - Research and Development - Machine Learning for Automation

Description

The company is a large industrial research organization building intelligent automation for manufacturing. This summer internship focuses on machine learning for robotic systems that can learn, adapt, and carry out complex manipulation, processing, and assembly tasks in real production settings, aiming for more flexible and efficient factories.

The role sits within a research team and suits someone who combines strong machine learning theory with hands-on work on physical robots and real-world problems.

Responsibilities

• Help develop and apply machine learning methods for robots, including reinforcement learning, imitation learning and learning from demonstration, and deep learning for perception, planning, and control.
• Apply learning-based methods to hard robotic tasks such as grasping and dexterous manipulation, mobile manipulation with coordinated motion, and contact-rich assembly with force-sensitive operations.
• Contribute to the team's technical work through presentations, intellectual property, technical documentation, and publications.

Requirements

• Required: enrolled in a Master's or PhD program in machine learning, physical robotics, computer science, or a closely related field at a leading university.
• Required: a strong academic and applied machine learning background, including experience training policies for grasping, manipulation, contact-rich assembly, or other robotic skills.
• Required: strong programming and system development skills in Python, C++, and PyTorch or TensorFlow.
• Required: you must currently be a student.
• Required: you must be able to work full time, 40 hours per week, during the summer.
• Required: you must be a U.S. Person, meaning a U.S. citizen, U.S. permanent resident, asylee, or refugee, because the role is subject to export control restrictions.
• Preferred: experience with dexterous and multi-contact manipulation.
• Preferred: experience with mobile manipulation in dynamic or unstructured environments.
• Preferred: experience with learning-based control for contact-rich assembly.
• Preferred: experience with simulation-to-real transfer and data-efficient learning.
• Preferred: experience with foundation-model-enabled robotics, such as vision-language-action models, transformer-based control, world models, and generative AI for robot planning and execution.
• Preferred: experience in industrial or automotive manufacturing environments.
• Preferred: a publication record in leading conferences such as ICRA, RSS, CoRL, or NeurIPS.

Conditions

This is a hybrid role. You are expected to report to the Warren, MI office at least three times per week, or at a frequency the business sets.

Start dates are May 24, 2027 or June 14, 2027.

The company does not provide immigration sponsorship for this role, now or in the future. Do not apply if you will need sponsorship or any work authorization that requires a written submission or other immigration support from the company, such as H-1B, OPT, STEM OPT, CPT, TN, or J-1.

Applicants may be required to complete a role-related assessment and/or a pre-employment screening before starting.

The company is an equal opportunity employer and does not discriminate unlawfully.

Benefits

• Eligible students selected for the 2027 Student Program receive a one-time lump sum taxable stipend to help cover relocation.