Raven finance -Smart contracts based under-collateralized P2P lending and borrowing

Name of Project: - Raven finance

Proposal in one sentence: -
A smart contract based p2p crowdfunding and risk mitigated borrowing system ,with various tiers based on the credit worthiness of the borrowers.

Description of the project and what problem it is solving:-
In Defi, there are majorly Collateralized and over-collateralized loans , due to the nature of decentralized and anonymous nature of web3; but Decentralized Identifiers ,web3 based credit systems, and users on chain reputations can be leveraged to provide an lending and borrowing ecosystem.

And here, we can build smart-contract based pools and lending
where lenders can earn much better interest rates than savings accounts
and FD’s ,

These kind of platforms do exists in web2 markets, but being on chain increases transparency, which often raises concerns for the investors.
Now here AI can be used to analyze the user profile and it’s on chain activities to determine the interest rates and risk management, matching the user’s profile.
- The global P2P lending market size was valued at USD 82,300 million in 2021, and is expected to touch USD 804,200 million by 2030, growing at a 29.1% CAGR.

Grant Deliverables:

  • Grant Deliverable 1
    Design the product, and research based upon the technical, legal, development aspects , and modelling of the infra on how the product will function.
  • Grant Deliverable 2
    Analysis of the AI tool based on the results, from the operations performed on a set of data, to determine, the borrowing capability to of the users wallets based and the interest rates at which the loans can be provided,

Squad

Divyansh

  • @Divyansh__001 -Twitter handle
  • DC #6011 - discord handle

Cool concept @divyanshh – think this is a really interesting area

What kind of decentralized identifiers do you plan to use? There are some tradeoffs between tying identity to an EOA wallet address vs something transferable like a safe or ENS domain.

Comparison against overcollateralized models like aave would be interesting here. In theory a user with zero credentials/credit still needs to be overcollateralized to start, but their rates and minimum ratios could decrease over time

Are you looking to actually train a machine learning model in this project? Or more focused on exploring the data available and prior work that’s out there.

Very interesting use case for AI and I would love to chat with you more about this project as I have some experience in this field.

I can see how it could be a significant help in solving the challenges of mutual credit risk assessment (especially for undercollateralized loans), ongoing credit scoring, and monitoring the overall health of the network.