Yishu Miao

Yishu Miao is the Co-founder and CEO of Haiper, an AI startup specialising in training Perceptual Foundation Models for visual content generation. Haiper is backed by prominent global investment funds and comprises a team of leading machine learning scientists and engineers from institutions like DeepMind, Google Brain, TikTok, and Meta. The company is developing a cutting-edge content creation platform.

Miao holds a Master’s degree in Data Mining from Tsinghua University, where he worked under Professor Chunping Li. His academic and professional experiences have equipped him with a deep understanding of machine learning, generative models, and multimodal systems. 

His research led to several notable publications, including “A Generative Framework for Simultaneous Machine Translation” (EMNLP 2021) and “Discovering Topics in Long-tailed Corpora with Causal Intervention” (ACL Findings 2021). Miao also held significant roles in the tech industry, including as Tech Lead Manager for Global Trust & Safety at TikTok and as a Postdoctoral Research Associate at Imperial College London, where he worked on Large Multi-modal Models with Prof. Lucia Specia.

Before his tenure at TikTok and Imperial College, Miao founded MO Intelligence, a startup dedicated to developing robotics technologies for the construction industry. His earlier career included research internships at DeepMind and contributions to the academic community, with publications such as “Cross-Modal Generative Augmentation for Visual Question Answering” (BMVC 2021) and “TextPlace: Visual Place Recognition and Topological Localization Through Reading Scene Texts” (ICCV 2019).

Miao’s academic background also includes a Master’s degree in Data Mining from Tsinghua University. His extensive research and leadership experience in machine learning, generative models, and multimodal systems underscore his expertise in advancing AI technologies for various applications.

Yishu Miao is the Co-founder and CEO of Haiper, a prominent AI startup focused on developing Perceptual Foundation Models for visual content generation. Backed by some of the world’s largest investment funds, Haiper is at the forefront of next-generation content creation technology. The company’s team includes top machine learning scientists and engineers from renowned organisations such as DeepMind, Google Brain, TikTok, and Meta, as well as prestigious academic research labs.

Before establishing Haiper, Miao completed a Doctor of Philosophy (PhD) in Machine Learning at the University of Oxford, under the guidance of Professor Phil Blunsom. His doctoral research, which centred on Generative Models for Language, led to influential publications, including “A Generative Framework for Simultaneous Machine Translation” (EMNLP 2021) and “Discovering Topics in Long-tailed Corpora with Causal Intervention” (ACL Findings 2021). These works reflect his significant contributions to the fields of machine learning and natural language processing.

Miao's professional trajectory includes a notable role as Tech Lead Manager for Global Trust & Safety at TikTok in London, where he led initiatives in machine learning and trust and safety measures. Additionally, he served as a Postdoctoral Research Associate at Imperial College London, collaborating with Professor Lucia Specia on developing Large Multi-modal Models. His work at Imperial contributed to the advancement of multimodal AI systems.

Earlier in his career, Miao founded MO Intelligence, a startup dedicated to developing robotics technologies for the construction industry. His experience also includes research internships at DeepMind, where he gained valuable insights into cutting-edge AI research.

Miao holds a Master’s degree in Data Mining from Tsinghua University, where he worked under Professor Chunping Li. His academic and professional experiences have equipped him with a deep understanding of machine learning, generative models, and multimodal systems.

In addition to his entrepreneurial and research pursuits, Miao has published several significant papers, such as “Cross-Modal Generative Augmentation for Visual Question Answering” (BMVC 2021) and “TextPlace: Visual Place Recognition and Topological Localization Through Reading Scene Texts” (ICCV 2019). His contributions to these areas demonstrate his expertise and leadership in advancing AI technology.

Yishu Miao envisions a future where AI-driven technologies revolutionise content creation and multimodal systems, enhancing the ways in which we generate and interact with visual and linguistic content. He aims to lead advancements in perceptual foundation models, making sophisticated AI accessible and transformative across various industries.