Responsibilities 

• Conduct cutting-edge research in retrieval and recommendation algorithms and multi-modality modeling with focus on developing machine learning (ML) models for understanding shortform videos using various modalities, including video frames, text descriptions, and background audio that are fundamental to moderation, searching, and recommendation for short-form videos. 
• Work on Data-driven training (with hyperparameters search) of the neural network models. 
• Perform offline evaluation (pseudo-A/B test) of the trained models using real-world short-form videos and document each model's detailed specification, training procedure, and evaluation result. 
• Communicate with the internal content moderation team on the technical needs (specifications and resources) to develop software for neural network models. 
• Review and analyze latest literature on the state-of-the-art neural network models for video understanding. 

• Publish research papers that benefit the academic community. 

   

Qualifications 

Must have a PhD in Computer Science, Engineering (any), Information Technology, Machine Learning, Data Science, Mathematics, Statistics, or a related quantitative field, and 1 year of related work and/or research experience. 

Of the required experience, must have 1 year of experience in each of the following: 

• Using data analysis tools SQL, PySpark, NumPy, and Pandas to handle large-scale datasets; 
• Programming using Python and PyTorch to build large-scale learning systems; 
• Developing data-driven prediction models using classic statistical methods, machine learning, and neural network/deep learning; 
• Signal processing for image, video, and audio data preprocessing; and 

• Preparing technical documentation, including technical proposals, patents, or publishing conference or journal papers. 

   

Type: Full time, 40 hours/week 

Location: Bellevue, WA 

Salary Range: $152214 - $237500 per year 

To apply, and for information on benefits offered, visit: https://jobs.bytedance.com/en/{rel="nofollow"} and type Job ID A158315A in search bar. Contact lpresumes@bytedance.com if you have difficulty applying through our website.