Software Engineer @ – Implementing efficient digital signal processing algorithms in C for a revolutionary small cell backhaul solution in mobile wireless networks. From July 2012 to Present (3 years 4 months) Berkeley, CAGraduate Student Researcher @ – Worked in Berkeley Parallelization Laboratory (Par Lab) with Prof. Kurt Keutzer.
– Parallelized a state-of-the-art natural language parser (BerkeleyParser) using CUDA. Tailor-designed efficient algorithms for GPU architectures to achieve a 25x speedup without loss of parsing accuracy.
– Parallelized a high-quality human recognizer with pose detections using Matlab/OpenMP/CUDA and achieved a real-time speed.
– Parallelized a machine translation decoder based on dynamic-programming algorithms using CUDA with a 4.3x speedup without loss of translation quality. From July 2009 to May 2012 (2 years 11 months) Berkeley, CAResearch Intern @ Designed algorithms to parallelize/distribute a machine translation training pipeline on a cluster of 40+ nodes using Hadoop MapReduce/Python/Perl with up to 10x speedup and also a slight gain in model quality. From June 2011 to August 2011 (3 months) Yorktown Heights, NYResearch Intern @ On incorporating the computational knowledge engine Wolfram|Alpha into the search engine Bing. From June 2010 to August 2010 (3 months) Redmond, WA
Master of Science (MS), Electrical Engineering and Computer Science, 3.96 @ University of California, Berkeley From 2008 to 2012 Bachelor of Science (BS), Electrical Engineering, 4.0 @ National Taiwan University From 2003 to 2007 Chao-Yue Lai is skilled in: Programming, Electrical Engineering, Algorithms, Machine Learning, Machine Translation, Natural Language..., Parallel Programming, EDA, C++, C, Java, Python, Perl, CUDA, Hadoop, LaTeX, Distributed Systems