San Francisco-based deeptech startup Antimattr is developing a voice-first hardware interface designed to facilitate human-agent interaction in noisy environments. Founded in 2025 by Sridipto Ghosh and Sirsha Chakraborty, the company is currently participating in the Y Combinator accelerator program, becoming the first India-origin wearable company to receive backing from the organization.
Technology and Product Development
The startup aims to solve the problem of background noise hindering voice AI adoption. According to CTO Sirsha Chakraborty, current voice interfaces struggle to function accurately in public spaces like cafes or crowded areas. Antimattr’s primary project, dubbed Project Mnemosyne, utilizes a hardware system comprising a ring and specialized earphones.
The system is designed to isolate the user’s voice from ambient sounds, allowing for clear dictation or interaction with AI agents regardless of the surrounding environment. CEO Sridipto Ghosh stated that the technology allows users to communicate privately and accurately in loud settings. The company credits its early development phase to the Scaler Innovation Lab, which provided access to prototyping tools such as 3D printers and soldering stations.
Strategic Direction and Industry Focus
The name “Antimattr” reflects the founders’ philosophy of challenging conventional industry notions. Despite receiving initial industry advice against entering the hardware sector due to competition from larger firms, the founders proceeded with their development plans. Participation in Y Combinator has accelerated the startup’s ability to refine its approach to human-machine interaction on a global scale.
Ghosh noted that the company’s evolution has been iterative, emphasizing that startups do not need complete solutions before beginning development. The company continues to refine its hardware interface to improve usability for voice-first applications.