In recent years, India's online food delivery and home services market has grown rapidly. While giants like Zomato and Swiggy have gone public, platforms like Urban Company and Pronto have digitized the workforce for daily tasks. Now, Silicon Valley-based startup Human Archive is leveraging this trend to collect first-person (egocentric) video data needed to train robots, reports Techcrunch.com reports .
Human Archive equips workers with camera-mounted caps to record their daily tasks. The startup currently partners with companies in home services, hostels, and restaurant sectors, having deployed over 1,000 active devices. Through this project, the company raised $8.2 million from Wing Venture Capital, NVP Capital, Y Combinator, and angel investors from tech giants like OpenAI, NVIDIA, Google, and Meta.
The company was founded by students from the University of California, Berkeley and Stanford University — Samay Mani, Rushil Agarwal, Shloke Patel, and Raj Patel. All have academic backgrounds in robotics, hardware, and tactile data. The startup's main goal is to solve the key problem facing AI labs and robotics companies: the shortage of high-quality data reflecting real-world physical actions.
However, not all companies welcomed the idea. For instance, major platforms like Urban Company and Pronto declined to partner with Human Archive. While Urban Company head Abhiraj Singh Bhal stated they would not enter such agreements, Human Archive founders called this decision a mistake that could lead to losing customers in the future. Nevertheless, India's emerging gig economy remains the largest "training ground" for robots.
Human Archive does not intend to stop at video. They are developing special gloves that accurately record movement and tactile force, full-body motion capture suits, and wrist cameras. The company believes that combining video data with sensor feedback will allow robots to be trained more accurately and efficiently.
Read “Zamin” on Telegram!
