Research Intern - Video Data Compression and Application

About the Team Our Multimedia Lab aims to explore and lead cutting-edge technologies in the multimedia field, deeply participate in international multimedia standardization work, provide software and hardware solutions for multimedia content generation, analysis/processing, innovative interaction, etc., and empower our business in many aspects, such as cost, experience, and capabilities. We are seeking exceptional Ph.D. students with research backgrounds in video and image compression, Gaussian splatting-based coding, token compression, and large multimodal models (VLMs/LLMs). Ideal candidates should demonstrate a strong track record of research excellence, strong capabilities in large multimodal model applications, a passion for advancing the future of multimedia technologies, and the potential to become leading experts in the field. Internships at ByteDance aim to provide students with hands-on experience in developing fundamental skills and exploring potential career paths. A vibrant blend of social events and enriching development workshops will be available for you to explore. Here, you will utilize your knowledge in real-world scenarios while laying a strong foundation for personal and professional growth. It runs for 12 weeks. PhD internships at ByteDance provide students with the opportunity to actively contribute to our products and research, and to the organization's future plans and emerging technologies. Our dynamic internship experience blends hands-on learning, enriching community-building and development events, and collaboration with industry experts. Applications will be reviewed on a rolling basis - we encourage you to apply early. Please state your availability clearly in your resume (Start date, End date). Responsibilities - Design, develop, and optimize innovative algorithms for data compression, processing, and large multimodal model applications, including but not limited to 2D video, multiview video, point clouds, Gaussian splatting–based or NN-based coding, token compression, and KV cache optimization. - Stay up to date with state-of-the-art techniques through standardization activities or leading conference/journal publications. - Build prototypes and demonstrations, and contribute to technical reports, publications, and patent filings. Minimum Qualifications - Current Ph.D. student in computer science/electrical engineering/mathematics/statistics/data science and related disciplines - Strong Computer Science fundamentals (algorithms, data structures, software design) and problem-solving skills - Familiar with token/image/video coding and processing or large multimodal models - Familiar with Python, PyTorch - Familiar with C/C++ - Collaborative mindset, with solid written and verbal communication skill Preferred Qualifications - Good understanding of state-of-art compression algorithms - Rich experience and interest in video/image coding standards (e.g., H.263/264/265/266, MPEG-2/4, JPEG, JPEG 2000, AV1, AV2, AVS1/2/3, etc.) - Proficient in deep learning frameworks such as PyTorch, TensorFlow, and YOLO. - Hands-on experience with large language models (LLMs) and vision-language models (VLMs), including LoRA, diffusion models, and VQA tasks. - Experience with model training, fine-tuning, and evaluation pipelines. - Familiar with GPU acceleration and CUDA environment configuration.