Tink Onboards Customers in a Couple of Days Using AWS

Open-banking platform Tink builds, trains, and deploys machine learning models in hours and onboards customers in a couple of days using AWS. The company provides a platform to banks, fintechs, and startups, which allows them to offer personal-finance-management tools and other services. Tink runs its platform on AWS and has introduced Amazon SageMaker to improve the efficiency of using machine learning.

Stockholm-based Tink is a leading European open-banking platform that enables banks, fintechs, and startups to develop data-driven financial services. Under the European Union’s second Payment Services Directive (PSD2), banks must allow third-party providers (TPPs), which include fintechs and startups, to access account holders’ transaction data through open application programing interfaces (APIs).

Through one API, Tink’s clients can access aggregated and enriched financial data, initiate payments, and build digital banking experiences, allowing them to gain new customers and retain existing ones.

Tink’s platform runs on the Amazon Web Services (AWS) Cloud. Mattias Appelgren, technology lead for infrastructure at Tink, says, “We chose AWS because of its managed services, so that instead of hiring an entire infrastructure team to maintain our IT services, we could focus on delivering value to our customers.”

Using AWS, Tink provides banks, fintechs, and startups with a dedicated environment on the Tink platform in a couple of days—a process that could take weeks or months with an on- premises environment. Appelgren says the speed comes from the ease of use of AWS, and he gives the example of working with the Amazon Relational Database Service (Amazon RDS). “You don’t need database expertise to use Amazon RDS. Our developers simply write code and connect to an Amazon RDS instance. Working this way streamlines onboarding and services development.”

To improve performance, Tink uses AWS Lambda, a serverless computing service, and has written Lambda functions to enforce policies for improving the utilization of its Amazon Elastic Compute Cloud (Amazon EC2) instances.

Tink uses Amazon SageMaker to train machine learning (ML) models that categorize the payment transactions that are aggregated from banks in Europe. SageMaker allows Tink to experiment with ML models that can reveal new patterns in bank transaction data to improve categorization services for customers.

Appelgren says, “Using Amazon SageMaker, Tink saves time training its ML models. In hours, engineers can build, train, and deploy ML models that reach millions of end users.”