Description:

Predictive features transformed the digital landscape in Web2, driving innovations like Amazon’s product recommendations, Google’s personalized searches, and Visa’s fraud detection mechanisms. Now, as Web3 emerges, it begins to leverage similar technologies for token and contract recommendations, personalized NFT curation, and the identification of airdrop farming activities. Yet, the complexity of building predictive features in the decentralized, multifaceted Web3 environment, with its multiple chains, millions of transactions, and addresses, is a formidable challenge. That’s where Seshat shines!Core contributors of Seshat are successful academics with a PhD in computer science and over 30 years of experience in data science and machine learning, numerous publications in top-tier ML and data science venues, and a successful track record of launching and delivering products.Born out of an extensive research project at the University of Waterloo, Seshat was developed through significant efforts to understand and address the pain points in predictive analytics for Web3. Seshat is deeply rooted in research and tailored to meet the real-world needs of developers in the Web3 ecosystem.Our product offers a range of services that includes an engine to build training datasets, an inference engine with low latency, a real-time data provider for inference, and off-the-shelf predictive features including vectorized Web3 entities (such as vectorized public addresses that can be used for downstream tasks).Developers will save 80% of time and resources when building predictive features with Seshat.