My journey into the world of building technology began in 2014 when I taught myself electronics to solve a problem I faced as a teenager spending a lot of time outdoors playing cricket in Bangalore. The problem was of my phone losing charge and avoiding the burden of carrying a brick with me every time I went out. So I built a power-sharing device that wirelessly shared juice between mobile phones. It stuck to the back of the phones like a sticker, with a small circuit and a coil connected to your usb port, and when you bring your phones together, one would get charged from the other. I sold this to a few of my friends and we actually used it charge our phones.

This experience compelled my dive into building technology at a creative capacity, and I spent the next 4 years tinkering with various AI and hardware technologies that shaped my perspectives during my undergrad in Bangalore, before I started my first company in 2018.

Here are some notable explorations before I started my first business:

1. Developing a clinical AI system to automate the diagnosis of vascular dysfunctions in Indian clinics. Publication
2. Building a BCI to correlate EEG activity with task performance using representation learning.
3. Building speech based deep learning pipelines on FPGAs in partnership with Intel, which was later used for Indian Defense deployments.

Now, jumping to 2018. As machine learning engineers fresh out of college, we started an ML services business to solve real-world problems with AI. We landed Target as our first client and developed a predictive maintenance solution for the devices used by staff in their 1000+ stores. This project helped us stay bootstrapped. But in the process, we discovered the challenges that AI projects face before they can take off, and one big problem was data sharing and exchange between data-owning teams and data-using teams. We learned that privacy, ownership, and transparency were the primary problems that surfaced in many different forms across different types of organizations.

We started a research lab and started experimenting with many privacy-preserving ML approaches like, federated learning, Trusted Execution/Confidential Computing, Multiparty-computation, differential-privacy and a few more. We learnt that CC was particularly fitting given their ability to scale more efficiently for the conditions we observed in certain markets. We were part of Intel, NetApp, Microsoft, and Cisco’s accelerators to refine and develop the technology to prototype new products that addressed data-sharing related challenges in large infrastructure enterprises.

We raised 2M from Accel in 2020 to build the market. After 3 micro-pivots and multiple iterations of customer discovery, we landed on Lending Fintech as a beachhead market with many pilots but unfortunately failed to scale the business due to the impact of the end of ZIRP period which impacted the American fintech market drastically (and many other reasons in hindsight). But this experience was rife with crucial learnings about customer discovery and understanding nuances in market structures.