Research Scientist Graduate - Multimedia Lab

We are looking for talented individuals to join our team. As a graduate, you will get opportunities to pursue bold ideas, tackle complex challenges, and unlock limitless growth. Successful candidates must be able to commit to an onboarding date by the end of the year. Please state your availability and graduation date clearly in your resume. About the Team At Multimedia Lab, we push the boundaries of what’s possible in multimedia technology. Our mission is to pioneer cutting-edge research across image and video understanding, generation, processing, compression, and transmission—and transform these innovations into real-world products that delight hundreds of millions of users globally. The ideal candidate combines deep technical expertise with a strong record of innovation, thrives on solving challenging problems at scale, and is passionate about shaping the future of multimedia experiences. This is an opportunity to work alongside top talent, drive frontier research, and turn breakthrough ideas into impactful technologies used around the world. Topic Content: Multimodal Foundation Models for Intelligent Multimedia Processing Explore next-generation multimedia technologies powered by multimodal foundation models, including perceptual quality modeling, generative enhancement, temporal video understanding, user-centric evaluation, and intelligent visual representation/compression, to advance video quality, efficiency, and user experience in future multimedia systems. Challenges for the analysis, understanding, and quality assessment and enhancement based on multimodal large models: - Modeling complex time sequences in long multimodal videos - Building few-shot grounding-based models for quality assessment - Creating interactive video processing/enhancement models aligned with user preferences Research value of analysis, understanding, and quality assessment and enhancement based on multimodal large models: - Enhance semantic understanding and event localization in medium- and long-length videos, improving processing efficiency, and support key areas such as ads recommendation, content comprehension, video value evaluation, and transcoding enhancement - Lower the cost of quality annotation, enable interpretable assessment of local degradation, boost generalization across different content types, and support pixel-level quality inspection and optimization Responsibilities - Design video analysis (ROI/SOD, content understanding, temporal grounding etc.) and quality assessment algorithms, and participate in database creation, algorithm design/development/optimization, etc. - Participate in designing strategy and solution for E2E video quality optimization with a combination of video analysis, processing and encoding algorithms - Apply designed algorithms for VOD / Live streaming monitoring, data analysis, objective evaluation for algorithms etc. - Collaborate with cross-functional teams to integrate algorithms into production workflows and validate their impact through A/B testing. Minimum qualifications: - Individuals who are completing or recently completed a PhD in Software Development, Computer Science, Computer Engineering, or a related technical discipline. - In-depth knowledge of video analysis algorithms or subjective/objective video quality algorithms, and state-of-the-art technologies - Proficient in one of the following: C, C++, Python Preferred Qualifications - Familiar with ML and image processing tools, including sklearn, opencv, ffmpeg, etc - Familiar with deep learning frameworks (Tensorflow/Pytorch) - Familiar with Transformer architectures and mainstream multi-modal large models (MLLMs), and hands-on implementation or research experience preferred. - Familiar with Linux development environments, shell scripting, HDFS etc - Knowledge of common video processing algorithms, such as supperresolution, defusion model, etc. - Great communication, eager to learn, and always passionate about turning cutting-edge technologies into real life use cases.