Title:Object Detection And 3D Pose Estimation - Computer Vision Expert
Posted: January 8, 2019
Company/Institution: Mujin Inc.
Location: Tokyo, Japan
Department: Perception group
We gather the smartest people from all over the world to tackle the most challenging robot automation problems, in order to deliver the latest and best robot technologies into factories, warehouses, food centers, and many others.
We build real-time vision and motion planning solutions that have a huge demand in industrial settings. We are looking for expert vision engineers to develop object recognition and 3D pose estimation solutions in scenarios that require real-time robot perception such as random bin-picking.
You will be part of the core R&D team within MUJIN and will focus on the algorithmic design, development and deployment of the computer vision applications. We work in a fast paced, agile environment, where we constantly collaborate, peer review ideas, and test proposed solutions. We encourage continuous learning and brainstorming of ideas within the team. You will need to:
- Solve cutting-edge scientific and technical challenges related to recognition and pose estimation of a very wide variety of objects in challenging scenarios.
- Analyze and evaluate state of the art algorithms related to detection and pose estimation, from academic and industrial research, and implement and enhance them in a production environment.
- Design, develop, and evaluate the performance of highly scalable and accurate real-time computer vision applications, in Python and C/C++.
- Perform detailed test, carry out performance analysis on algorithms and use continuous integration tools to finely enhance the rapidly deployed algorithms.
- PhD in Computer Vision related topics or MSc with vast computer vision experience.
- Expert in 3D pose estimation of textured and textureless objects in cluttered scenes, with proven experience.
- Very solid Python and C++ experience.
- Experience with a vast set of computer vision libraries.
- Solid mathematical background.
- Previous contributions to open source projects.
- Advanced Linux user.
- Extras: object tracking, SLAM, computer graphics, augmented reality, machine learning, exposure to projects in robotics.
- MSc Degree or higher in computer science or related fields.
- 3+ years of academic or work experience in object recognition or pose estimation.
- High proficiency in Python and/or C++.
- Experience in creating computationally efficient code that generalizes to multiple detectors.
- Ability to turn academic ideas into high-quality code.
- Excellent knowledge in linear algebra.
- Highly enthusiastic about robotics.