A new artificial intelligence startup named Mirai has secured $10 million in seed funding. The company, founded by the creators behind popular apps Reface and Prisma, aims to develop technology that improves how AI models run directly on consumer devices such as smartphones and laptops.
The investment round was announced this week. The funds will be used to advance Mirai’s core mission of enhancing on-device model inference, which refers to the process where an AI algorithm makes predictions or generates content without needing to communicate with a remote server.
Background of the Founders
The co-founders bringing Mirai to life are Denys Dovhan and Yaroslav Boiko. They previously founded AI-powered face-swap application Reface. They are joined by Alexey Moiseenkov, the original founder of the AI photo-editing app Prisma. Their combined experience in creating consumer applications that utilize neural networks positions them at the intersection of AI research and mainstream product development.
Their prior ventures, Reface and Prisma, gained widespread popularity for implementing complex AI features like style transfer and face animation directly on users’ devices. This background in efficient, on-device AI processing is a foundational element for the new company’s direction.
The Technical Challenge and Goal
Currently, many advanced AI models require significant computational power, often provided by large data centers in the cloud. This setup can lead to latency, dependency on internet connectivity, and potential privacy concerns as user data is sent to remote servers for processing.
Mirai’s stated objective is to overcome these limitations by making sophisticated AI models run more efficiently on local hardware. Improving on-device inference could lead to faster response times, greater user privacy, and reduced operational costs for developers by minimizing cloud computing expenses.
The company has not disclosed specific technical details about its approach. However, the field generally involves techniques like model optimization, compression, and hardware-aware neural architecture design to shrink large models so they can operate within the memory and power constraints of a mobile device.
Market Context and Implications
The push for capable on-device AI aligns with broader industry trends. Major technology firms like Apple, Google, and Qualcomm are heavily investing in chip design and software frameworks to enable more AI processing locally. This shift is driven by demands for real-time applications in areas like photography, translation, and voice assistants, where latency is critical.
For consumers, successful advancements in this area could mean more responsive and private AI features in everyday apps. For developers, it could lower the barrier to integrating powerful AI into applications without incurring high infrastructure costs or compromising user experience.
The $10 million seed round indicates significant investor confidence in the founding team’s vision and technical capability. The capital will primarily fund research, development, and team expansion as the startup works to translate its research into tangible products or developer tools.
Looking Ahead
Mirai is expected to begin scaling its engineering team in the coming months. While the company has not announced a specific product launch timeline, its work will contribute to the growing ecosystem of tools aimed at decentralizing AI processing from the cloud to the edge.
Industry observers will be watching for technical publications, partnerships with hardware manufacturers, or eventual SDK releases that allow other developers to leverage Mirai’s optimization technologies. The long-term success of the venture will depend on its ability to deliver measurable improvements in speed, efficiency, and model performance on a wide range of consumer devices.
Source: Various
