Search button in the site


Machine Learning/Signal Processing for Sensor Data, Research Engineer

Position Summary We are working on several exciting technologies to estimate 3D position and orientation of persons and objects using a small, lightweight, and inexpensive IMU(accelerometer/gyroscope). Taking full advantage of its characteristics, notably not being affected by the surrounding environment, we aim to develop a robust motion sensing technology by optimally combining deep learning and sensor fusion techniques.
We are looking for research engineers with great passion to drive innovation for applications such as robot control and content creation in entertainment industry in addition to electronics. Specifically, those who are interested in the following areas are most welcomed: IMU Device, Indoor/Outdoor Positioning, Motion Capture, or Motion Retrieval. Our projects cover sensing applications as well as sensing devices, and our mission is to bring differentiation to next generation Sony products.
Responsibilities Develop machine learning or signal processing algorithm, and implement it as a software library or digital circuit.
Repeat problem setting, data collection, algorithm improvement, and field test to bring it to a practical technology.
Design optimal software and hardware, and make a working prototype to demonstrate its customer value.
Propose new user experience proactively while surveying academic papers and industry trends.
Work in a team, also typically with Sony business units or external partners, whether domestic or overseas.
Required qualifications Must be familiar with coding in C/C++, Python, Rust, Android/iOS, MATLAB, or machine learning frameworks.
Good to have experience with:
- Unity/Unreal Engine, Motion Builder, Maya, or any other CG software.
- FPGA, PSoC, Raspberry Pi/Arduino, or any other electronics.
Preferred qualifications Good to have knowledge of:
- MEMS IMU and inertial navigation system.
- GNSS or radio-based positioning.
- various sensor fusion techniques such as Kalman filter.
- SOTA deep learning techniques.
Even better to be familiar with:
- sensor fusion processor.
- wearable sensing device.
- time series data processing.
- data aggregation via crowd sourcing.
Product, Service Digital camera, Mobile robot/drone, RTLS, AR navigation, XR interaction, Game and video content production, Posture measurement for sports and rehabilitation.
Development Environment Windows/Linux/Mac, Embedded/Cloud.
Essay Required
Coding Test Not Required
Location Tokyo, Japan
Page Top