Research Scientist (Privacy-Preserving Machine Learning)

Machine Learning, Computer Vision

Full Time

Location flexible (Tokyo, Zurich, US)

General Information

Sony AI is seeking research scientists to join our Privacy-Preserving Machine Learning (PPML) team. Our team mainly focuses on the fundamental and applied research in CV privacy. The role of a research scientist is to conduct innovative research, develop methodologies, design and prototype solutions, and collaborate with engineering teams to empower real-world AI applications (e.g., Imaging&Sensing) in Sony.

Roles and Responsibilities

• Conduct fundamental and innovative research in privacy and security related areas, including but not limited to data anonymization, differential privacy, adversarial training, machine unlearning, homomorphic encryption, federated learning, secure multi-party computation, training/inference on encrypted data, etc.
• Deep understanding of trustworthy AI, privacy and security threats in various CV applications, such as Image classification, object detection, pose detection, semantic image segmentation, etc.
• Identify new privacy-preserving and robust machine learning approaches and rigorously assess their validities on resource-limited end/edge devices, such as Sony IMX500/501.
• Design and prototype secure and privacy-preserving system to empower real-world AI applications in Sony, specifically Imaging&Sensing business.
• Build long-term collaboration with AI, Engineering and Product teams, seek for high-impact opportunities to launch solutions and advance long-term business growths.
• Publish influential research outcomes, give presentations in top-tier conferences and journals.

Required Qualifications and Skills

• Work experience with a role with primary emphasis on AI privacy and security.
• PhD (or equivalent experience) in Computer Science, Electrical Engineering, Math, Statistics, data science, or other related disciplines.
• Publications or expertise in privacy and security, computer vision, tiny-device learning, adversarial learning, distributed/decentralized system design, communication efficiency, etc.
• Familiar with Python, C++, Pytorch, Tensorflow, etc.
• Excellent communication and presentation skills.
• (Preferred) Experience in leading or participating in significant research projects.
• (Preferred) Paper awards in top conferences, or winner of international competition award.

Working Location

Location flexible (Tokyo, Zurich, US)