Summary
Hongbo Xu is a Machine Learning Architecture Engineer based in Beijing with eight years’ experience optimizing large-scale deep learning systems at Baidu Search. He specializes in model compression and distributed training for LLMs, routinely applying quantization and pruning to cut inference cost while preserving accuracy. His work spans production inference frameworks (brpc + Paddle Inference) and PaddleSlim-based INT8 quantization pipelines, enabling high-throughput deployment of advanced models. Trained at Sichuan University and the University of Chinese Academy of Sciences, he combines solid academic foundations with practical, system-level engineering that balances performance, cost, and reliability. A concise motto—Per aspera ad astra—reflects his persistence in turning research techniques into robust production solutions.
8 years of coding experience
Bachelor of Engineering, Computer Science and Technology, Bachelor of Engineering, Computer Science and Technology at Sichuan University
Master's of Engineering, Computer Science and Technology, Master's of Engineering, Computer Science and Technology at University of Chinese Academy of Sciences