Artificial Intelligence for Video Compression - Research Scientist
Core
Develop next-generation neural video coding technologies and contribute to future video compression standards.
Role type
Research Scientist (Learning-Based Video Compression)
Builds
Neural video coding algorithms, learned motion representations, and entropy models for multimedia systems
Domain
Multimedia engineering, machine learning, video compression standards
Deliverable
production ML models
Required skills
Machine learning algorithms, video coding theory, rate-distortion optimization, Python, C/C++, PyTorch
Preferred skills
Publications in leading conferences, patent applications, experience with VVC/H.266, HEVC/H.265, AV1
Responsibilities
Conceive, develop, implement, and optimize learning-based video compression algorithms; Evaluate and improve neural/hybrid compression systems; Represent Qualcomm in standardization forums (JVET, MPEG Video, ITU-T/VCEG); Document and present new algorithms via papers, patents, and presentations
Seniority
Mid-to-Senior level (2+ years for PhD, 3+ for Master's, 4+ for Bachelor's)
