Research Scientist in Cambridge

At the Samsung AI Center-Cambridge, our teams work closely to develop innovative machine learning (ML) algorithms and systems, with the overarching objective to deliver state-of-the-art efficient AI models across Samsung consumer devices. We focus on high-impact projects that balance research and commercial outcomes. Our group conducts research on on-device and edge-based ML systems, focusing on model/system and hardware/software co-design, adaptive inference methods, robotic systems, and the optimized on-device deployment of large language models (LLMs) and multimodal visual language models (VLMs). Our team is characterised by its diverse backgrounds and expertise, ranging from pure machine learning and mobile/embedded systems to computer architecture and robotics, with a strong emphasis on combining excellent research skills with hands-on development abilities. This diversity has enabled us to develop groundbreaking cross-stack ML systems and algorithms. Throughout our projects, we maintain strong ties with close-to-product teams, contributing to commercialization goals and ensuring our research has a tangible impact for Samsung through successful tech transfers and numerous patent applications. Our work is consistently and widely recognised, with multiple papers published in top-tier international conferences and journals.

Role and Responsibilities

• Research and development of world models that enable AI agents to understand, simulate, and reason about complex environments.
• Design and run experiments at the frontier of model-based learning, contributing to foundational research that shapes how AI agents perceive and interact with the world.
• This role requires a strong background in machine learning, a passion for tackling open-ended research problems, and the drive to build systems.
• Contribute to research and commercialization efforts for Samsung AI Centre in areas including on-device LLM and VLMs, adaptive inference methods, and mobile ML systems.
• Conduct cutting-edge research within the existing group's agenda as well as contribute towards shaping it.
• Collaborate closely with cross-functional product teams, as well as on-site research teams, to integrate ML solutions into consumer devices.
• At the core of our efforts is innovation, as you will design groundbreaking machine learning algorithms and systems that extend the capabilities of current technology.
• You will translate research findings into practical applications, contributing to the commercialization of AI technologies across millions of Samsung devices.
• Additionally, you will prepare comprehensive documentation, research papers, and contribute towards patent applications.
• You will work alongside team members with varying levels of experience in a collaborative and continuous learning environment, with a strong focus on individual development throughout our work.
• In our group, we highly value and require an open and cooperative mindset and seek determined individuals who are committed to obtaining new knowledge and producing high-quality work.

Skills and Qualifications

Education and experience:

• PhD in CS/EE or related research experience in academia or industry.
• We will consider various levels of experience in relevant research areas.

Key Skills:

• Experience with ML frameworks (PyTorch, TensorFlow, JAX) and efficient ML (incl. quantization, pruning, sparsification, distillation, etc.).
• Experience with deployment on embedded/mobile devices (such as smartphones, with mobile CPU, GPU, NPU).
• Experience with distributed and multi-GPU training at scale.
• Fluency in Python, C/C++ and GNU Linux.
• Proficiency in code version control, Git and GitHub.
• Experience in working as a member of a team.
• Solid publication record of papers in top-tier venues, such as NeurIPS/ICLR/ICML/MobiCom/MobiSys/ICCAD/MLSys.

Any of the following skills will also be positively considered:

• Experience in real-world mobile system deployment.
• Research experience in efficient Generative AI, including language, visual or multimodal tasks. This includes different aspects of the pipeline, from data and preprocessing to large model adaptation, fine-tuning and on-device optimization.
• Android operating system and Android app development.

Samsung has a strict policy on trade secrets. In applying to Samsung and progressing through the recruitment process, you must not disclose any trade secrets of a previous employer.