PhD Research Intern, AI for Robotic Manipulation

San Francisco

Chef Robotics

Chef Robotics | Empowering Humans to Do What They Do Best

View all jobs at Chef Robotics

Apply now Apply later

Chef is looking for PhD interns to help us develop cutting-edge, AI-based manipulation methods for our fleet of robots. You will work with a research mentor to conduct research on novel, learning-based approaches to robotic manipulation of food, and evaluate results on real robotic systems used for food production in a manufacturing environment. This is an exciting opportunity to apply the latest AI research to a real-life practical application, and leave a lasting impact on Chef and the food production industry.

Responsibilities and Example Projects

  • Develop DL-based computer vision, motion planning, and/or manipulation algorithms that work across thousands of different foods with varying physical properties (stickiness, density, viscosity, etc.).
  • Develop algorithms to learn ingredient properties from past data or on-the-fly and intelligently adapt manipulation parameters.
  • Incorporate new sensor data (e.g. force control, additional cameras) into learning algorithms.
  • Contribute to software infrastructure for training, deploying, and evaluating ML models.

Requirements

  • Currently enrolled in a relevant PhD program focused on embodied AI or robotics.
  • Experience with at least one of the following areas: imitation learning/learning from demonstration, reinforcement learning, or multi-modal learning.
  • Able to independently drive and execute on a well-scoped but open-ended research project.
  • Experience with common deep learning frameworks like Tensorflow or Pytorch.
  • Excellent programming skills.
  • Excellent mathematical and statistical foundation in ML concepts.
Apply now Apply later
Job stats:  0  0  0

Tags: Computer Vision Deep Learning Machine Learning ML models PhD PyTorch Reinforcement Learning Research Robotics Statistics TensorFlow

Region: North America
Country: United States

More jobs like this