@wugi00

Deep Learning Engineer

SK commnunications•  July 2017 - Present

= Developed a deep learning-based end-to-end speech synthesis system. ◦ Developed a TTS system using open sources and added Korean handling features, modifying some of those sources using Python and Tensorflow. ◦ Changed the guidance voice of the actual mobile phone navigation app to a synthesized voice. ◦ Architected and programmed batch scripts and web pages for collecting data, preprocessing and post-processing using Python, bash, MySQL, PHP, ffmpeg. ◦ Some synthesis audio samples here. https://wugi88.github.io/Tacotron-WaveNet-samples/ = Developed a personalized recommendation system. ◦ Developed a user age gender prediction model using Keras.. ◦ Developed a user clustering program for news recommendation based on their activity using Python, Pandas, scikit-learn. ◦ Developed a personalized recommendation system by Side feature added Matrix Factorization for News and BBS sites using Tensorflow, Python, HDFS.