Principal Scientist, Reinforcement Learning

Key Skills

Computer Vision (CV), Confluence, GitLab, JIRA, Leadership, Machine learning techniques, Python Programming, PyTorch, Reinforcement Learning, TensorFlow

Job Description

Your New Role and Team

Sanctuary is seeking an exceptional Principal Scientist, Reinforcement Learning (RL) to join our team in engineering and innovating unique robotic manipulation tasks.

As a Principal Scientist, you'll be responsible for selecting the most promising state of the art (SOTA) approaches, designing training and data-collection pipelines, supervising the process of testing these algorithms in simulation, and deploying them to our robots in real-world settings. With access to in-house robots, you'll also have a unique opportunity for impactful work with new haptic and proprioceptive sensing modalities.

Success Criteria

• Design, implement, and improve SOTA RL algorithms and test them in real-world settings
• Keep up to date with SOTA RL methodologies and robotics
• Identify, communicate, and drive promising research directions to the team
• Find ways of improving existing implementations of RL learning pipelines with regards to standard metrics such as sample efficiency, speed, computational resource usage, and scalability
• Design RL training and data-collection pipelines to facilitate fast deployment on physical robots
• Work with a multidisciplinary team to develop novel algorithms and investigate sources of errors with existing implementations

Your Experience

• QualificationsPhD in Machine Learning, Computer Science, Applied Math, or equivalent work experience in ML
• 4+ years' experience implementing a variety of ML methods with a focus in a specialization such as computer vision or robotics
• 2+ years' experience implementing and deploying (dexterous) robotic manipulation tasks in simulation and on physical robots
• Experience in simulation-to-reality transfer learning
• Experience taking ML R&D and trained models into production
• Hands-on experience integrating ML models onto a robotics platform 
• Experience with computer vision systems
• Publications in leading AI conferences such as Neurips, ICLR, ICML and CORL

Skills

• Development with Python 3.6 or later
• Working knowledge of PyTorch and/or TensorFlow
• Familiarity with ROS2
• Extensive knowledge of Reinforcement Learning principles and use
• Experience with Atlassian tools; Jira, Confluence, or equivalent i.e. GitLab
• Traits Above all else, a consistently positive attitude and a willingness to do whatever it takes to create robust solutions to complex problems
• Strong leadership skills in organizing R&D work for ML projects
• Eager to take on new challenges with tenacity and positivity
• Patience, persistence, and attention to detail when resolving performance issues
• Enthusiasm for bringing human-like intelligence to machines
• Ability to drive development of new functionalities from concept to production
• Ability to multitask and prioritize in a fast paced environment.

Working at Sanctuary

Sanctuary is an equal opportunity employer; employment with Sanctuary is governed based on skills, competence, and qualifications and will not be influenced in any way by race, color, religion, gender, national origin/ethnicity, veteran status, disability status, age, sexual orientation, gender identity, marital status, mental or physical disability, or any other legally protected status.

Benefits

Full time (non co-op) employees enjoy medical/dental/vision coverage, life insurance, wellness programs, stock options, paid time off (3 weeks vacation accrued annually, 12 paid holidays, 5 days of annual sick leave, and parental leave), scheduling and worksite flexibility by role, and more.

About Sanctuary

Founded in 2018 by Geordie Rose, Suzanne Gildert, Olivia Norton, and Ajay Agrawal, Sanctuary is a Vancouver, Canada-based company. Sanctuary is on a mission to create the world’s first human-like intelligence in general-purpose robots that will help us work more safely, efficiently, and sustainably. And in the not-too-distant future, help us explore, settle, and prosper in outer space.

Members of the Sanctuary team founded D-Wave (a pioneer in the quantum computing industry), Kindred (first use of reinforcement learning in a production robot), and the Creative Destruction Lab (pioneered a revolutionary method for the commercialization of science for the betterment of humankind). The team has experience launching market-defining innovations rooted in previously unsolved and deep scientific problems.