Duke University and VaultLink Inc: Shaping the Digital Asset Standards for Banks

A Partnership for Innovation

Duke University’s Pratt School of Engineering in Financial Technology Program collaborates with VaultLink Inc. (“VaultLink”) to lead FinTech innovation, which is dedicated to the fast-growing and intricate realm of blockchain and traditional finance (“TradFi”). VaultLink offers an innovative SaaS solution for banks to provide digital asset services in a regulatory-compliant manner. This marginally banked sector is currently valued at over $2.5 trillion. Recent regulatory changes, exemplified by the successful bitcoin ETF launches by industry giants like Blackrock and Fidelity, have increased the need for a standard to support our financial institutions in servicing digital assets.

James Slazas, CEO of VaultLink., shared: “Duke University’s exceptional students brought incredible creativity to a challenging and virtually uncharted problem: how to apply the same protections and controls on digital assets in our financial system as we do with traditionally banked assets. Our financial institutions need this support as we shift from moving fiat to being a digital value transfer rail for all tokenized assets.”

This partnership led to the creation of an advanced risk-scoring model to help standardize the assessment of digital assets to be transferred into the banking system safely. The project employed machine learning and blockchain transaction data analysis to forge a cutting-edge risk-scoring model, focusing on adding real-time surveillance tools for KYC (Know Your Customer) and layer of protection with KYA (Know Your Asset). KYA tracks the provenance of assets and the correlation between other wallets and a customer wallet to ensure bad actors have not utilized the assets. The aim was to equip banks to safely accept and transfer digital assets to the same control standards as fiat in the banking system. This initiative marks a critical step toward bridging the gap between innovative FinTech solutions and regulatory banking standards.

The Vanguard of Innovation: Student Contributions

Twelve students, guided by Dr. Alessio Brini, a Postdoctoral Researcher researching for the Digital Asset Research & Engineering Collaborative (DAREC) lab, and Jacob Vestal, Executive in Residence in the Engineering Graduate and Professional Programs at Duke University’s Pratt School of Engineering, applied data analysis tools and innovative machine learning models to bolster their research.

Fangrui Huang, Ziyan Zhang, Wanglin Cai, Chungsong Ma, Zhiyuan Chen, and Kunpeng Zhu extensively explored the current landscape in blockchain analytics. They focused on risk typologies and their implications for Anti-Money Laundering (AML) practices, scrutinizing traditional and emerging risk assessment methodologies and metrics. Their research and insights finally concluded qualitative and quantitative risk factors that were fundamental in the following machine learning model development process.

Rebecca Xiao, Siyuan Fan, Qian Lu, Zhui Jian, Siqi Yu, and Zichen Zhang queried large amounts of raw blockchain data, laying the groundwork for comprehensive data preparation. Utilizing Python, they used exploratory data analysis to reveal statistical correlations between transaction attributes and associated risk typologies (or risk cohorts). Their efforts led to the development of sophisticated machine learning algorithms, both supervised and unsupervised. While supervised models leverage historical risk-labeled data to achieve higher predictive accuracy, unsupervised models revealed novel patterns and correlations, enriching the comprehension of risk factors inherent in blockchain transactions. The combination of model strategies advanced the precision of risk evaluation and facilitated the exploration of more sophisticated risk prediction models. The models were designed to enhance risk assessment, which is essential to ensure that only trustworthy assets enter banks without association with malicious activities.

"It is genuinely motivating to observe our students not merely engaging with these practical challenges but also succeeding in developing advanced solutions that contribute significantly to the advancement of the financial technology sector."

Our collaboration with Vault Link was truly inspiring. The comprehensive introduction to Blockchain and its business applications provided by the VaultLink team and Prof. Brini was both specific and thorough. Thanks to their profound expertise and wealth of experience, we completed the model construction with a high-quality standard,” said Qian Lu, one of the participating students.

Another student, Rebecca Xiao, reflected: “The project offered me profound insights into the seamless integration of emerging FinTech with traditional financial systems. Digging into the blockchain transaction data and engineering potential risk features required combining our team’s creativity and analytical tools. Understanding the transaction activity patterns hidden in data allowed us to develop exciting risk model frameworks.”

Dr. Alessio Brini expressed his appreciation for the students' contributions to the project, stating, "I am impressed by the commitment exhibited by our students in addressing the complex issues associated with the blockchain domain. This initiative demanded a comprehensive array of skills, integrating both data analysis and machine learning—essential competencies underscored within our Financial Technology curriculum. It is genuinely motivating to observe our students not merely engaging with these practical challenges but also succeeding in developing advanced solutions that contribute significantly to the advancement of the financial technology sector."

Through this project, the students researched existing risk assessment frameworks, delved into blockchain transaction data, and developed innovative machine learning models for digital asset risk assessment. Duke University’s Pratt School of Engineering and VaultLink laid the groundwork for establishing standards to ensure the safety of digital assets entering the banking system.