News and resources on cloud strategy, selection, build, migration and operation for banks and fintechs.
NatWest Markets eyes big gains in migration to Google Cloud

NatWest Markets eyes big gains in migration to Google Cloud

As it transitions to Google Cloud, NatWest Markets is aiming to build artificial intelligence and machine learning capabilities that can detect financial risks for customers before they arise.

Faced with the need to process larger volumes of data at speed and on ageing servers, the investment arm of the UK bank committed to migrate all of its risk platforms to the cloud in 2019.

To build cloud capabilities within the company, the firm invested in training 1,100 employees on Google Cloud Platform throughout 2021, with almost 100 of them becoming Google Cloud certified.

The first step in its journey into the cloud entailed moving data processing workloads to Google's BigQuery engine in order to cope with escalating volumes of data.

Chris Conway, head of risk and finance technology at the bank, says: “By simply migrating from on-premise to BigQuery, we’ve already noticed a 60% improvement in compute time for overnight batch processing of risk simulations and calculations."

The bank is using multiple Google Cloud products to achieve these outcomes including Compute Engine, Google Kubernetes Engine, Cloud SQL, Cloud Storage, BigQuery and Dataproc.

With its migration journey well underway, NatWest Markets is looking to future operational gains in regulatory reporting and in the use of AI to to predict future financial shocks and business opportunities for customers before they arise.

“Our goal is to have all risk modelling, surveillance, and supervision powered by machine learning on Google Cloud from 2022 onwards, where we can better connect the dots from different data sources,” says Conway.

NatWest Markets is also looking at how Google Kubernetes Engine can help improve its performance by containerising application components as it transitions from a monolithic to a microservices architecture. “We aim to speed up our release cycles from 12 months to 2 weeks for the XVA (valuation adjustments) desks, and we’re just touching the surface of what we’re able to do with Google Cloud,” says Conway.

Comments: (0)