Last Time Again
- Venice Biennale College CinemaSelected
Thomas Tang was born on 28 December 1999 in rural Guizhou, China, and now lives in Beijing.
A graduate of Beijing Film Academy and Peking University, he is currently pursuing a doctorate.
He has won awards in every category of the Xianli Awards, the highest student honours of Beijing Film Academy’s Cinematography Department, and was named an Outstanding Graduate of Peking University. Films he has helped create as a key crew member have been selected for programmes at the Venice Biennale, Sundance, the Student Academy Awards, the Shanghai International Film Festival, the Golden Rooster Film Festival, the Beijing International Film Festival and FIRST, among other leading festivals. He has also received more than twenty individual honours.
During his postgraduate studies, he turned his attention to AI-generated content and cross-media practice, exploring how emerging technologies might shape the future of filmmaking. He has served as a video-model training expert and product advisor for companies including ByteDance, Alibaba and SenseTime.
As product lead, he is now working with his team on an AI agent for professional filmmakers. The product has become an industry benchmark in video understanding and the evaluation of cinematic generation. The company has secured an eight-figure seed round, and multiple research outputs have been adopted by more than 100 AI labs, including Google DeepMind, ByteDance and xAI.
Film & productMaking a film and designing a creative tool both begin with understanding intent. I translate filmmaking experience into practical products that help creators retain their judgment, control and voice when working with generative AI.
Translating the practical demands of filmmaking into product decisions, creative workflows and model evaluation standards.
Everyday observations in street, portrait and landscape, on film and digital, arranged by mood.
Drag to explore · click to view
Festival selections, project programmes and awards for films I helped create.
Directing, cinematography and editing across AI films, documentaries, experimental work, VR and theatrical features.
Research on bringing professional filmmaking knowledge into the evaluation of generative AI.
Co-author · Mingshuang Tang (Thomas Tang)
A film-grade benchmark for text-to-video and reference-to-video generation grounded in professional cinematic language. It combines multi-shot prompts derived from real films, a hierarchical evaluation framework and automated evaluation tools to assess cinematic craft.
Programs, scripts and small apps I wrote to solve concrete problems in creative work.
The things I keep doing outside work; each one opens to its own page.
Articles and interviews on film, product and technology.