Bounty

Table 50

Jihoon Yang (31D) Tony Yeon (293) Terry Kim (388)

Inspiration

From AI research experience, our team members have faced financial burden by the lack of compute resources and we wished to make it more equitable.

What it does

Our system, built on top of Solana’s blockchain network, allows users to upload untrained ML models along with an escrowed bounty to incentivize others with significant compute power to invest their resources for model training.

Our system is designed such that when an individual invests their compute, they will receive a fraction of future profits for when the model goes live in prod. With this feature, we hope to build an investment incentive for all involved parties to create mutual benefit for those in web3 as well as those in AI.

How we built it

The three of us split the work between Solana development, Next.JS back end development, and React front end development. Throughout the hackathon we switched roles often so that we could spend time leaning the full stack together.

By leveraging the decentralized validation of the Solana chain as well as the raw compute output of Google Cloud, we believe we’ve built a potential solution to the issue compute equity.

Challenges we ran into

Since we are new to blockchain development, we found aspects of using the Solana chain rather difficult.

Accomplishments that we're proud of

However, by doing heavy research into Solana and how users interact with accounts and programs, we were able to get a fledgling understanding of Solana and build our own unique blockchain solution.