Dr. Yassine Maalej is a researcher who has spent the past decade at the intersection of large-scale video pipelines and applied AI. Much of that work was done at Spectrum (Charter Communications), which helped shape how video is processed, understood, and monetized at hyperscale. He architected burst transcoding of multi-camera replay streams on NVIDIA NVENC across H.264, HEVC, and AV1, instrumented VoD player quality metrics under hardware impairment, and built the computer-vision and language-model systems behind Spectrum's CDN video analytics and delivery platform — including visual context summarization with vision-language models, CLIP-based hybrid video search, and LLM-driven contextual video understanding.

His foundation in artificial intelligence and video processing comes from his Ph.D. in Electrical and Computer Engineering at the University of Idaho, where his research centered on deep learning and 3D perception — image object detection and tracking, video semantic segmentation, point-cloud alignment, and real-time video processing using vision frameworks like Caffe, Darknet, OpenCV, ROS, and CUDA.

Earlier, as a research assistant at Pennsylvania State University, Yassine worked on hardware design for camera systems, developing automated Verilog/VHDL workflows to optimize circuits for power, area, and timing. That hardware-level fluency, paired with his vision research and production video experience, gives him an unusually full-stack view of video encoding and delivery systems.

Yassine's work spans both industry invention and published research. He currently holds 16 granted US patents and academic research publications with over 200 citations spanning generative video, semantic video analysis, video streaming delivery, and LLM-driven video analytics. 

​