Research Scientist Internship

Sports AI Tech Lab develops AI technology for understanding sports from game video, player and ball tracking, event detection, and tactical indicators. This research role applies language, speech, and multimodal methods to viewing experiences in services such as ABEMA and WINTICKET, linking research outputs to sports media products.

Work may include generating commentary, explanations, highlights, titles, or descriptions from video, play, tracking, and event data; building Japanese streaming language models; speech synthesis and low-latency audio; real-time translation; user-adapted recommendations; evaluation of generated language and audio; and real-time prediction of match events or player performance. The posting names large language models, multimodal generative AI, machine learning, and Python implementation.

Applicants must have doctoral-level expertise and research experience in at least one of natural language processing, speech processing, dialogue systems, information recommendation, or multimodal machine learning. They must also have implementation experience using Python for machine-learning models and large language models.