Use a Machine Learning model to optimize yield

I posted this on Discord as a question a while back, and was recommended to make a thread here on the forum. My understanding of Sperax is that it is not specializing in the business of optimizing yield, i.e that’s a secondary concern to growing the use of USDs, hence it would make sense to leverage a protocol that would specialize in optimizing the yield of the treasury ie the stablecoins backing USDs.

There are multiple projects that are leveraging machine learning to optimize yields, the one I am the most familiar with is Mozaic Finance, they not only have trained their models for over a year on on-chain data, but also enable cross-chain yield optimization via LayerZero. I could provide more details about their models and optimization strategy, but I would like to just open up the conversation to see if this something of interest to the community at large, and also the viability of doing that within the boundaries and plans for the existing roadmap for 2024.


I am owner of the xMOZ for a while at the moment, and I think that their AI-tech will really become one of the major competitors to Yearn Finance. Both projects have pretty much similar aim, however Mozaic Finance (Arbitrum based btw.) utilizes their novel AI system to search for the best yield on-chain. I do believe Sperax could use their tech to further improve yield for USDs, and getting even better returns.


Yield optimisations using ML sounds like a great idea. I think if a great proposal comes in, the Sperax DAO should incentivise protocols building on top of it. Could you ask the protocols to come up with a proposal for it?


Hey, really appreciate the feedback, glad to get sone positive replies to this. I will share that with the Mozaic team, and hopefully they’ll come up with one.

1 Like

Indeed we will - this sounds like great opportunities to synergise here!

1 Like


I think you should have intriguing idea to leverage Mozaic Finance’s machine learning models for optimizing yields to enhance the utility and efficiency of Sperax’s USDs treasury. Given that Sperax’s primary focus is on growing the use of USDs integrating a specialized yield optimization protocol could be beneficial. This approach could potentially maximize returns without detracting from Sperax’s main objectives. Opening a discussion on this topic within the community and evaluating its alignment with the 2024 roadmap would be a prudent next step.