Robotics Automation Perception Engineer
Tokyo
Automation – Automation - Retail /
Full-time /Part-time /Contract (English) /
Hybrid
This role focuses on automation of custom made manipulators, aiming to automate convenience store operations without human intervention to a high success rate, by developing perception modules.
The focus will be to leverage senor data (RGB data, Depth data, and others) provided from operational robots in real world, and improve the existing perception system to further automate its success rate. In addition, this role also focuses on developing a new perception algorithms that can further boost its performance. Having those focuses given, optimizing / compressing deployed algorithms to minimize limited system usage will be also required. We strongly believe that our robots would provide great impact to the current market and further one, as well as would greatly be happy seeing others feeling the same.
Responsibilities
- Contribute to the development of an automation system using custom hardware, by developing sophisticated software for perception system, with deep learning, computer vision, sometimes combined with other control algorithms.
- Work with other automation engineers and achieve priorities and goals set by CxOs .
- Break up vague goals into concrete steps and manage your own timeline to achieve company goals.
- Accurately analyze the problem and data, evaluate and deploy the system to the product with safety and stability taken into account, within given timelines.
- Collect data, analyze results, and continuously improve algorithms quantitatively
Requirements
- 3+ year of experience of developing perception system or equivalent (including research field)
- Quick and accurate problem solving skills
- Careful issue monitoring / debugging skills
- Strong experience / understanding in low level computer vision, machine learning, and deep learning, as well as experience of implementing state-of-the-art methods from scratch.
- Proficiency in Python, and libraries commonly used for deep learning and computer vision such as : pytorch (and other wrapped libraries such as mmdetection, fastai) tensorboard, mlflow, wandb, opencv, open3d, point cloud library
- Experience in image sensors with RGB and depth information (including separate sensors combined such as RGB camera + LiDAR )
- Experience in domain randomization using OpenGL, Blender, Isaac Sim, MuJoCo etc.
- Strong experience in making customized version of any of the following deep learning applications: object detection / segmentation and key point detection. 3d pose estimation and 3d reconstruction
- Experience in performance optimization : network layer optimization and TensorRT / TorchScript etc
Preferred Skills
- Experience in coding with modern C++ (C++11, 14 and above)
- Experience with ROS
- Knowledge about camera calibration / hand-eye calibration.
- Research experience in CV / AI and related robotics field (CVPR, ICRA and such)
- Work experience with manipulators
- Work experience in startups
- Engineering experiences with one or more of the following
- Motion planning and execution
- Signal processing
- network communication (pub/sub, req/rep, etc)
Position Location will be at Telexistence's Headquarters Office (Harumi, Chuo-ku, Tokyo) or another location designated by the company.
The role assignment may be modified based on organizational requirements. This could include changes to your job responsibilities, reporting structure, or team assignment.
