TECHNICAL GRANT APPLICATION – DeRisk - TIER 3

#1

TERMS OF USE

  • YES, I have read and understand the CONFLUX ECOSYSTEM GRANTS OVERVIEW and confirm that my proposal is suitable for the Growth Grants stream.
  • YES, I understand that the Conflux Growth Grants are subject to a No Sale Rule and have read its terms.
  • YES, I acknowledge that I will be required to provide additional KYC information to the Conflux Foundation to receive this grant.
  • YES, I understand and agree to follow public grant reporting requirements.

APPLICATION INTRODUCTION
DeRisk is another venture lead by the team of the successful project Carmine Finance, a project taking care of DeFi risks, including an Automated Market Maker enabling anyone to buy and sell options at a fair price (mostly traders, liquidity providers and Hedge Funds and investors) or insurance against crypto price drops.
DeRisk itself was conceptualized to address the inherent risks in the DeFi lending sector. As an open-source platform, DeRisk focuses on monitoring and forecasting the risk of loans, primarily those at risk of going under the water, to protect lending protocols and their users. By providing an “early warning” system, DeRisk aims to enhance capital efficiency and security within the DeFi ecosystem.

DeRisk operates on the principle of integrating with lending protocols and AMMs analyzing loans against available capital and liquidity risks. The platform has been developed with a keen understanding of the volatility in the crypto market and the need for efficient capital allocation.
The goal of DeRisk is to provide a robust and scalable solution that reduces the chances of bad debt accumulation in the ecosystem, improves the confidence of users in lending protocols, and encourages more capital flow, thereby enhancing the overall health of the DeFi ecosystem. Furthermore, thanks to the increased transparency, the lending protocol could be more open to adding mid to mid-low level tokens.

Webpage: DeRisk Official Site
DeRisk App Demo: DeRisk on StarkNet

TEAM

  • Team Leader: Marek Hauzr (LinkedIn) with extensive experience in Finance, HFT, AI/ML, MEV. Responsible for project management and strategic direction.

  • Team members: Ondrej (LinkedIn), Andrej (LinkedIn), David (LinkedIn), and Guillermo (LinkedIn) with diverse expertise in development, finance, and blockchain technology, contributing to various aspects of DeRisk’s development and operations (more in the Team section in Carmine Finance webpage)

TOTAL BUDGET AND FUNDING TIER
Grant Tier: Tier 3
Grant Size: $36,000

  • $5k - Environment setup and customisation of current infrastructure
  • $20k - Integration of Lending Protocols
  • $6k - Integration of AMMs
  • $5k - Visualization (FrontEnd)

CURRENT FUNCTIONALITY
DeRisk currently has a working version built on StarkNet. Official release is planned for mid December. It is capable of monitoring lending protocols, comparing loans against available capital on DEXes and liquidators’ smart contracts, and detecting loans with higher risk.
To illustrate that in practice, consider Lending Protocol A and Lending Protocol B that have lent ~$5 mil. worth of tokens while available capital on DEXes is just above $15 mil. This situation poses a considerable risk not only to that protocol, but to other lending protocols and the broader DeFi ecosystem.

With DeRisk implemented, the 2 lending protocols would know of such a situation while it’s brewing, providing some previous time to react. They could:

  • Stop issuing loans that have too much risk, or adjust collateral needed. When they know a specific loan could go under water… They could also say ahead of time say “loans A to Z get additional incentive for liquidation” which would result in liquidators bringing more capital.
  • A lot less capital needs to be in security pools of lending protocols because of understanding the risks - meaning a lot more capital in circulation on Conflux
  • Be ready to buy insurance in the future (we provide the possibility of buying insurance as options as a hedge mechanism).

Additionally, DeRisk would help Conflux in more ways:

  • It standardizes loan data so that it is easier for liquidators to be onboarded not only as liquidators for a single protocol but for multiple ones with low additional cost.
  • It will also be possible to forecast probable arbitraging opportunities on a chain level, resulting in a higher market efficiency.
  • DeRisk will also contain analytics and insights into the lending protocols and compare them between each other. Giving their users more confidence in not only lending protocols but also in Conflux.

TECHNICAL PROPOSAL
The technical scope of DeRisk involves:

  • Infrastructure setup including data collection.
  • Integration with lending protocols and AMMs on Goledo, dForce, Swappi, Moon Swap, and MemeDex.
  • Development of a mechanism for automated search and integration of liquidators.
  • Implementation of a system for monitoring altcoins/utility tokens within the ecosystem and their liquidity risks.
  • Implementation and deployment of customized frontend.
  • Running the DeRisk for next 12 months.

SYSTEM MODEL
DeRisk utilizes a combination of Python, Streamlit, PostgreSQL, and React. This technology stack ensures robust data collection, processing, and a user-friendly interface. The system model includes data collection through an indexer, data standardization, and an open API for data access. More details can be found directly on the working version built on StarkNet.
Our solution offers further synergies such as insurance against bad loans. More on that can be found in a separate Doc.

DEVELOPMENT ROADMAP
The integration and development process for DeRisk is expected to take:

  • Milestone 1: Environment setup and customisation of current infrastructure (1 month)
  • Milestone 2: Integration of Lending Protocols (1-2 months)
  • Milestone 3: Integration of AMMs (1-2 months)
  • Milestone 4: Visualization (FrontEnd) (1 month)
  • Milestone 5: Running DeRisk (12 months)

MAINTENANCE CONSIDERATIONS
DeRisk plans for maintenance include:

  • Everything developed within the scope of this grant will be under a MIT license
  • Continuous support and running of the DeRisk application for 12 months.
  • Providing basic updates to the protocols as required.
  • Adding more lending protocols is complex and we would apply for an additional grant
2 Likes
#2

Hello @Nicolas_Fanta thanks for applying for a Conflux Grant.
Here are some questions I have about your proposal:

  1. Can you elaborate on the algorithms or models you use to assess and forecast the risk of loans?

  2. What are the key features of the user interface you are planning to develop?

  3. What strategies do you have in place to encourage adoption of DeRisk among lending protocols and their users?

  4. How do you plan to provide continuous support and updates for DeRisk over the 12-month period and beyond?

  5. What metrics or indicators will you use to measure the success and effectiveness of DeRisk?

  6. Beyond the 12-month timeline, what future developments or expansions do you foresee for DeRisk?

#3

1. Can you elaborate on the algorithms or models you use to assess and forecast the risk of loans?

  • To assess bad debt risk, rather than complex modeling we strive for giving the most accurate picture possible. That is that we gather all loans, see for which price level they will have to be liquidated, and based on simulated price changes for the specific AMMs, we estimate the liquidity available at the price level (agent based simulation method). To get the full picture we aggregate that across all AMMs. We then have benchmarks for first warnings (i.e. price levels concerning low liquidity levels).
  • Thanks to being able to monitor all loans on the lending protocols we then constantly have a very clear picture of the liquidity situation. This monitoring is however costly to implement, but its strength is its simplicity, transparency and constant visibility.
  • We understand that this might not be the easiest to understand and we would be delighted to explain that further on a call.

2. What are the key features of the user interface you are planning to develop?

  • Graphical representation of the liquidable debt and corresponding supply of liquidity to liquidate those loans if needed.
  • Call to actions for situations that present too much risk.
  • Lending protocol comparison on a chain level.
  • API to get access to the current loans.

3. What strategies do you have in place to encourage adoption of DeRisk among lending protocols and their users?

  • DeRisk will be freely available under a MIT license and it will also be running as a web-app. Lending protocols and their users will have that as a free feature helping them prevent bad debt. They will be able to set warning triggers in case the risk increases so that they can eventually implement preventive measures.
  • Our goal is also to talk to all the lending protocols on Conflux and convince them to use DeRisk. Part of this discussion will also be the implementation of the lending protocols to find potentially differences between deployed smart contracts and documentations.
  • Monitoring of the risk of bad debt is the key building block necessary to be able to build insurance against it for anyone. For insurance against bad debt we plan to leverage our current skills at Carmine where our goal is to take care of risk on chain (specifically standardized risk in DeFi) and discuss and implement this insurance.

4. How do you plan to provide continuous support and updates for DeRisk over the 12-month period and beyond?

  • We have a monitoring and a person allocated to make sure that DeRisk is up and running constantly.
  • Within the budget we will update minor changes on the lending protocols and AMMs. Integration of major updates and new lending protocols would have to be discussed with Conflux’ Grants Committee and we might apply for an additional grant as the implementation is challenging and resource intensive. There is an option to include some budget for these cases within this application but based on our experience “getting money for something that might or might not happen” is usually not well received. I would like to state that our goal is to make DeRisk used and that means we have to provide a valuable product. This means we want to support changes as much as possible.
  • All upgrades and improvements we will prepare on the infrastructure will be automatically added to the DeRisk on Conflux app.

5. What metrics or indicators will you use to measure the success and effectiveness of DeRisk?

  • We see our success indicator in the commitment of lending protocols to use DeRisk and use DeRisk’s output to influence their behavior.
  • Our goal for DeRisk is for it to be self-sustainable. So within the 12 months we want to provide such value for the lending protocols that it will be hard to replace DeRisk. Meaning that after the 12 months the lending protocols will support DeRisk.

6. Beyond the 12-month timeline, what future developments or expansions do you foresee for DeRisk?

  • We (or someone else) can naturally create his own BI or reporting on top of the data and we will do this. Including comparison of Conflux against other ecosystems and protocols within Conflux.
  • First we would like to integrate the feature of insurance against bad debt accumulation.
  • Next we would like to improve the user experience via integrating further triggers (depending on the feedback acquired within the first 12 months).
1 Like