Shawn Shen is an artificial intelligence researcher and entrepreneur, best known as the co-founder and Chief Executive Officer of the AI research lab Memories.ai. Before founding his company, Shen was an AI research scientist at Meta Reality Labs. [1] [2]
Shen holds a Ph.D. in Computer Science from the University of Cambridge. His academic background in computer science provided the foundation for his subsequent career in artificial intelligence research at major technology companies and later as an entrepreneur. [3] [1]
After completing his doctorate, Shen joined Meta as an AI research scientist at its Reality Labs division. During his time at Meta, he was involved in the company's research and development efforts in artificial intelligence. He credited his experience at Meta for helping shape his vision, noting its "bottom-up" innovation culture and the opportunity to collaborate with other researchers. However, Shen also expressed frustration with the frequent reorganizations within the company, which he felt could be a "waste of time" for researchers whose managers and goals were subject to constant change. This experience, coupled with a personal ambition to "start a great company," led to his departure from Meta in late 2024. [1]
In 2024, Shen co-founded Memories.ai with Ben Zhou, another former researcher from Meta Reality Labs. As CEO, Shen leads the company's technical vision and business strategy. The company was established to address what Shen identified as a critical gap in AI capabilities: long-term visual memory. Under his leadership, the company secured significant seed funding and began an aggressive talent acquisition campaign to compete with established technology giants for top AI researchers. [2]
Memories.ai is an AI research company founded by Shen and Ben Zhou in 2024. The company's stated mission is to develop artificial intelligence with human-like visual memory capabilities. [2]
The core innovation at Memories.ai is the development of what the company calls the world's first Large Visual Memory Model (LVMM). This technology is designed to solve a fundamental limitation in many AI systems, which can typically only process and recall information from short video clips before their context is lost. The LVMM enables an AI to capture, structure, and index vast amounts of video data, creating a persistent and searchable "visual memory layer." This allows the AI to understand context, identify recurring patterns, and detect changes over extended periods, effectively giving it a long-term memory for visual input. Shen described the product as "ChatGPT for your video library, with unlimited video context!" [3]
Shen has stated that while leading models like Google's Gemini were limited to analyzing approximately one hour of video, the context window for the LVMM is "virtually unlimited." The company also claims that its video understanding benchmarks demonstrate an "ultra-low hallucination rate." According to performance benchmarks shared by Shen in July 2025, Memories.ai achieved a score of 64.5% on MVBench, which was higher than scores reported for models from Google and OpenAI at the time. By September 2025, the platform had analyzed over 10 million hours of video to train its LVMM. [3] [2]
In a statement about the company's focus, Shen said, "For the first 3 years of this generative AI revolution, the majority of advancements and value has been created in text. Video is the most dense and entertaining format of information and we’re confident the next decade will see massive advancements with AI. We are working tirelessly to power that revolution." [2]
In the summer of 2025, Memories.ai announced it had raised an $8 million seed funding round. The round included participation from investors such as Susa Ventures, Samsung Next, Crane Venture Partners, Fusion Fund, Seedcamp, and Creator Ventures. This capital was raised to support the company's research and its efforts to recruit top-tier talent in the competitive AI landscape. [1] [2]
The company officially launched to the public on July 24, 2025. Shortly after, on July 28, 2025, it was featured as the #1 Product of the Day on the technology discovery platform Product Hunt. To encourage adoption and development on its platform, Memories.ai released its API to the public on August 4, 2025, allowing developers to build multi-modal AI agents that leverage its visual memory technology. [3]
In September 2025, Memories.ai launched an aggressive recruitment drive, publicly announcing that it was offering compensation packages of up to $2 million to attract leading AI researchers from major technology firms like Meta, Google, Microsoft, Anthropic, and xAI. Shen positioned this strategy as a direct response to the "talent war" in Silicon Valley, where large corporations were offering multi-million dollar packages to retain key personnel. He stated, "It's because of the talent war that was started by Mark Zuckerberg. When I heard about their compensation packages, I was shocked — it's really in the tens of millions range." [1]
The compensation packages were structured with a heavy emphasis on equity over cash, with Shen noting, "We are welcoming people who want to take more equity compared to cash... We will treat these hires as founding members, not as employees." The company's hiring goals were to bring on three to five researchers in the six months following the announcement, and another five to ten in the subsequent six months. This strategy proved successful in attracting high-profile talent, including the appointment of Chi-Hao (Eddy) Wu, a former Meta research scientist with over 13 years of experience in machine learning, as the company's Chief AI Officer on September 4, 2025. [1] [2]
As of September 2025, Shen was 28 years old and based in Cambridge, England. He has personally engaged in testing the applications of his company's technology. In August 2025, he revealed that he had been using a physical recording device for several months to capture his own life experiences. This footage was then processed by the Memories.ai platform to create a highly personalized AI assistant based on his own memories. [1] [3]