Responsibilities
1. Responsible for the research and development and business support of intelligent audio understanding and processing algorithms for ByteDance's audio content consumption business scenarios, including but not limited to Douyin, Xigua Video, live broadcast, clipping, etc. 2. Build system-level solutions for intelligent audio understanding and processing to provide technical firepower for ByteDance's intelligent audio ToB 3. Track the latest technological progress in the field of smart audio and upgrade various algorithm systems self-developed by the team, including 3D space rendering, audio quality improvement system, audio event detection system, audio understanding system, etc. 4. Follow up on the audio needs of the product business and continuously improve the audio quality experience of the product 5. Track and develop advanced audio progress in the industry, and develop and implement products in the speech/audio field with statistical model/machine learning/deep learning technology.
Qualifications
1. Rich experience in the development of digital signal processing and artificial intelligence/deep learning systems: project practice in one or more fields such as 3D spatial audio rendering, noise reduction/echo/dereverberation and other audio pre-processing, voiceprint/wake-up, sound event detection, speech recognition, natural language processing, etc. 2. Familiar with data structures and algorithms, deep network model design and tuning, proficient in open source tools such as Kaldi, TensorFlow, Pytorch, etc., experience in model training and exploration on large-scale training data sets is preferred 3. Good sense of teamwork and learning ability, business awareness, and enthusiasm for speech and audio field technologies 4. Priority will be given to those who have published papers in relevant international conferences or mainstream journals (ICASSP, Interspeech, ASRU).
