Search button in the site


Deep Learning, Researcher/Research Engineer/Software Engineer

Position Summary Our Deep Learning team in Sony R&D is building a software platform that accelerates and enhances the development of deep learning models dedicated for Sony's businesses such as entertainment businesses, including games, music, motion pictures, animation and other content intellectual property (IP); and electronics businesses including image sensors, digital imaging, audio & visual, mobile, and robots.
Our mission to make our platform unique and strong is 1) to develop novel fundamental techniques and reusable components of deep learning, and submit papers to top AI conferences to contribute the community and get feedback; 2) to explore new applications of emerging techniques in the deep learning community and to bridge the gap between research and production while closely communicating with business units, for examples, Sony Pictures Entertainment, Sony Interactive Entertainment, and Sony Electronics; 3) to develop core libraries and GUI tools, including Neural Network Libraries ( which is our own high performance deep learning framework which has broken the world record of a distributed training benchmark on ImageNet classification task and is provided as an open-source, and Neural Network Console (, a unique and professional-grade GUI for deep learning development.
Our team is located in Tokyo, Japan, but closely collaborating with diverse and talented teams in Germany, China, India and the United States.
Responsibilities Topics:
- Deep Learning General: Large-scale training & transfer learning, Generative models, AI-ethics technologies (XAI&Fairness), Architectures & training algorithms for specific data formats such as graph & cross-modal.
- Deep Learning Software Libraries: Development of opensource deep learning framework nnabla. (low-level implementation, API design, training reproduction of recent models, etc.)
- Creation AI core: CG and deep learning for games, movies etc.
Expected outcomes:
- Paper submission to top conferences.
- Publishing opensource software libraries/packages.
- Deep Learning models & tools (software libraries/packages or feasibility reports) for specific business applications at Sony.
Required qualifications Some of the following skills are required:
- Deep and wide understanding and knowledge of deep learning algorithms and software.
- Experience of designing and training deep learning models (almost) from scratch.
- Familiarity of signal processing, computer vision and computer graphics.
- Familiarity of generative models such as GANs, flow-based generative models, diffusion probabilistic models and so on.
- Familiarity of cross-modal (vision, language, audio etc.) deep learning or graph neural networks. (3D vision, relation data, etc.)
- Familiarity of AI-ethics and its techniques.
- Strong software engineering skills either in Python, C++, distributed computing, GPGPU, or API design.
Preferred qualifications Remarkable experiences:
- Strong research records in top AI conferences.
- Open source contributions.
- Whatever technical outputs. (books, talks, videos, blog posts, decks, etc.)
- Whatever professional engineering/research experiences. (intern, full-time, part-timer etc.)
Product, Service Electronics (On-device image/audio/sensor processing), Entertainment contents (Content management, Video/audio content enhancement), Games (Play Station), AI SaaS (Neural Network Console, Prediction One) etc.
Development Environment HPC servers (Linux), Python, C++.
Essay Not Required
Coding Test Required
Location Tokyo, Japan; Can work remotely depending the theme
Page Top