Role Overview
Qualcomm's Multimedia R&D seeks a Video Compression Research Engineer focused on learning-based video compression. Join a world-renowned team developing next-gen neural video coding technologies and contributing to future standards. Flexible global locations (USA, Germany, China, UK) with relocation support.
Responsibilities
- Develop novel learning-based video compression algorithms (neural video coding, learned motion representation, entropy modeling)
- Evaluate and improve neural/hybrid compression systems using ML techniques and rate-distortion optimization
- Represent Qualcomm in standardization forums (JVET, MPEG Video, ITU-T/VCEG)
- Document and present algorithms via standards contributions, patents, conference papers, and publications
Requirements
- PhD or Master's in EE, CS, ML, Applied Math or related field
- Experience developing ML models for compression, computer vision, multimedia
- Minimum: Bachelor's + 4 years, Master's + 3 years, or PhD + 2 years in Hardware/Software/Systems Engineering
- Innovative ML algorithms for learned image/video compression, computer vision, generative AI
- Knowledge of video coding theory, rate-distortion optimization, neural compression
- Experience with VVC/H.266, HEVC/H.265, AV1, neural video coding
- Track record of publications/patents at leading conferences
- Excellent Python and C++ programming, PyTorch experience