职位类型
技术-后端开发-算法工程师
职位描述
As a Speech Algorithm Engineer, you will:
Be responsible for speech data cleaning, text processing, data analyzing, etc.
Train automatic speech recognition models and speech evaluation models
Help to build and test the state-of-the-art speech recognition system.
Be responsible for speech data cleaning, text processing, data analyzing, etc.
Train automatic speech recognition models and speech evaluation models
Help to build and test the state-of-the-art speech recognition system.
职位要求
We hope you:
M.S. or Ph.D. in Signal Processing and Communications, Pattern Recognition, or Natural Language Processing (experience at high-level institutions and/or laboratories is preferred).
Have experience using Linux systems and are familiar with awk, sed and other tools
Have strong programming skills to implement algorithms (Good C/C++ programming and python/Matlab programming expertise is required).
Have solid learning and adapting capabilities; could collect and read English papers independently.
Are fluent in Chinese Mandarin
Candidates who can meet the following conditions are preferred:
a. experience with automatic speech recognition, voiceprint recognition, speech and language evaluation research; familiar with noise reduction algorithm and channel compensation algorithm.
b. experience with htk, kaldi, RWTH ASR, CUED-RNNLM and other open source tools; could potentially build complete system independently.
c. familiar with automatic speech recognition end-to-end framework (e.g. las, ctc etc.) and cnn, lstm, gru structures; familiar with discriminative training (discriminative/conditional models) and wfst based decoder.
M.S. or Ph.D. in Signal Processing and Communications, Pattern Recognition, or Natural Language Processing (experience at high-level institutions and/or laboratories is preferred).
Have experience using Linux systems and are familiar with awk, sed and other tools
Have strong programming skills to implement algorithms (Good C/C++ programming and python/Matlab programming expertise is required).
Have solid learning and adapting capabilities; could collect and read English papers independently.
Are fluent in Chinese Mandarin
Candidates who can meet the following conditions are preferred:
a. experience with automatic speech recognition, voiceprint recognition, speech and language evaluation research; familiar with noise reduction algorithm and channel compensation algorithm.
b. experience with htk, kaldi, RWTH ASR, CUED-RNNLM and other open source tools; could potentially build complete system independently.
c. familiar with automatic speech recognition end-to-end framework (e.g. las, ctc etc.) and cnn, lstm, gru structures; familiar with discriminative training (discriminative/conditional models) and wfst based decoder.
海外经验要求
海外留学经验优先
工作地点
北京市 - 北京市 - 北京市海淀区丹棱街6号丹棱soho12层