Guanhua Wang is a final-year Computer Science PhD student at UC Berkeley. His research lies primarily in the ML Systems area including fast collective communication schemes for model synchronization, efficient in-parallel model training and real-time model serving.