Alireza Heidar-Barghi

Senior Software Engineer @Tveon

Kelowna, BC, CA

Developed a framework to study FFmpeg threading structure.\\\\n• Identified tasks assigned to various types of threads and thread count in the video transcoding pipeline in FFmpeg.\\\\n• Explored the impact of idle threads on performance for FFmpeg-based encoders.\\\\n• Explored performance scaling by varying CPU core count on FFmpeg-based encoders.\\\\n• Investigated the feasiblity of improving multithreading performance for FFmpeg-based encoders.\\\\n• Investigated the feasiblity of enabling encoding parameter tuning at runtime for FFmpeg-based encoders.\\\\n• Benchmarked FFmpeg for various video transcoding use cases on different FFmpeg-based encoders.\\\\n• Automated time-consuming benchmarking and logging scripts.\\\\n• Documented multi-threaded application structure, benchmarking and profiling results on Confluence pages.\\\\n• Conducted profiling using Linux Perf, VTune, and Gperftools on various FFmpeg-based encoders (bare-metal, Docker container, and Kubernetes clusters).\\\\n• Benchmarked FFmpeg-based encoders in a batch processor as CI/CD jobs on a GitLab system.\\\\n• Investigated VMAF scores for some uses cases using FFmpeg\'s VMAF and Netflix\'s VMAF in various environments.\\\\n• Investigated VMAF scores using the easyVMAF tool for various use cases (scan mode, resolution, synchronization, and frame rate).\\\\n• Investigated banding artifact using Netflix\'s CAMBI tool and subjective assessment.\\\\n \\\\n\\\\n \\\\n \\\\n\\\\n \\\\n

Sharif University of Technology

Bachelor of Applied Science - BASc, Electrical Engineering

University of Toronto

Doctor of Philosophy - PhD, Electrical and Computer Engineering

Queen's University

Master's degree, Electrical and Computer Engineering