Built a digital bank from the ground up with gen AI integration
Runs 270 microservices consecutively
Reduces DevOps costs with BigQuery automation
Pave Bank, a startup aiming to modernize the financial services industry, chose Google Cloud to build a digital bank quickly. Using Google Cloud's Vertex AI, Pave Bank can accelerate product development.
As a startup aiming to disrupt the way people bank, Singapore and Georgian-based multi-asset and programmable bank Pave Bank built a bank integrated with generative AI capabilities using the Google Cloud Model Garden. The team believes in staying nimble and agile and does this by incorporating gen AI capabilities into its core infrastructure. “With Model Garden, we consume thousands of tokens a day directly from Google Cloud, over an API running in the data center where we are,” says Simon Vans-Colina, CTO and Co-founder at Pave Bank.
The team works to build a platform for customers to interact directly with their digital and traditional currency through using Vertex AI. “We are working from the ground up to incorporate AI directly into our core. Engineers also have direct access to the Vertex AI models so they can make decisions based on predictions without worrying about infrastructure or architecture,” adds Vans-Colina.
With Model Garden, we consume thousands of tokens a day directly from Google Cloud, over an API running in the data center where we are.
Simon Vans-Colina
CTO and Co-founder, Pave Bank
According to Vans-Colina, Google Cloud was the preferred cloud solution for Pave Bank due to its ability to seamlessly integrate tools like Cloud SQL and Dataflow with BigQuery. “It’s the fact that we could get access to the models that we needed through the Model Garden and Vertex AI,” he adds.
With tools like Pub/Sub and Cloud Run that enable the company to run stateless containers as well as Kubernetes, Vans-Colina says Google Cloud has made it easy for the team to build and run its platform.
Apart from efficiency, the team has also reaped the financial benefits of Google Cloud with BigQuery. “The automation capability of BigQuery means we don’t have to get our data scientists to do basic work like wrangling Python scripts to extract, transform, and load data and run analytics,” explains Vans-Colina. “Instead, we could trust that dataflow would ensure our changes are captured and replicated to BigQuery altogether, saving both time and money.”
The automation capability of BigQuery means we don’t have to get our data scientists to do basic work like wrangling Python scripts to extract, transform, and load data and run analytics. Instead, we could trust that dataflow would ensure our changes are captured and replicated to leave the basic scripting to BigQuery altogether, saving both time and money.
Simon Vans-Colina
CTO and Co-founder, Pave Bank
Pave Bank is reimagining how a bank is built, how it operates and how businesses interact with their bank. It is on a mission to give every business access to multi-asset, regulated, financial products. It is a technology company when it comes to product design and build, and a fully regulated bank when it comes to risk, capital and regulatory management.
Industry: Financial Services
Location: Singapore