onsite
Computer Vision Engineer, Data Monetization Technology
Computer Vision Engineer, Data Monetization Technology
As a Computer Vision Engineer in Data Monetization Technology at TikTok, you will be responsible for understanding business content across various platforms (ads, e-commerce, short video, live streaming) using images, text, video, and audio. This role involves data mining, feature engineering, and building robust machine learning models to optimize monetization ecology, along with exploring and implementing cutting-edge AI technologies on a large scale.
About the role
Responsibilities
- Be responsible for business content understanding of ads, e-commerce, short video, live streaming, and other related content understanding, including images, text, video, audio, etc.
- Be responsible for data mining, feature engineering, and building machine learning models to build monetization ecology.
- Optimize model computation efficiency and improve model stability when facing tens of millions of business data and restricted resources.
- Based on billion scale business data, explore and implement various cutting-edge technologies, such as pre-training, self-supervised learning, few-shot learning, etc.
Qualifications
- Bachelor's degree or above, majoring in Computer Science, Computer Engineering, Electrical Engineering, or other related fields.
- Have a solid foundation with common machine learning and deep learning related techniques and algorithms (e.g. classification, clustering, regression, etc.). Be proficient with at least one deep learning framework (e.g. PyTorch, TensorFlow).
- Be familiar with computer vision related tasks. Have rich experience in at least one aspect, such as image/video classification, object detection, image/video retrieval, OCR, image segmentation, etc.
- Related experience in at least one of the following areas is a plus:
- Be familiar with NLP-related tasks. Have experience in at least one aspect, such as text classification, semantic analysis, sentiment analysis, NER, etc.
- Be familiar with audio-related tasks. Have experience in at least one aspect, such as ASR, AED, LID, etc.
- Be familiar with multimodal machine learning, large-scale pre-training, etc.
- Be familiar with the theory and application of graph neural networks, knowledge graphs, and have relevant experience;
- Be familiar with model acceleration techniques such as pruning, quantization, distillation, etc.; Have relevant experience in deploying models using frameworks such as TensorRT.
- Solid programming foundation. Be familiar with basic data structures and algorithms.
- Have excellent analytical and problem-solving skills, logical thinking skills, communication and collaboration skills. Maintain curiosity about new things, and have a strong sense of responsibility, integrity and reliability.
- Having published papers in top AI conferences or journals is a plus.