How we scale our analytical capabilities with increasing demand from multiple clients
A talk by Avision Ho, Lead Data Scientist @NatWest Boxed
At NatWest Boxed, we provide Banking-as-a-Service (BaaS) to consumer brands and fintechs so they can offer financial products such as card accounts and flexible lending directly to their customers. The more businesses work with us, the more demand there is to understand the data we collect and process on their customers. However, how can a team of 10 data scientists and analysts scale flexibly and rapidly to meet this growing demand for data insights by more and more businesses working with NatWest Boxed?
In this talk, I will cover how we leverage Large Language Models (LLMs) to create a text-to-sql product that democratises access to our data so that we can provide quicker and easier insights. This way, we can flexibly scale to meet the growing demand for data insights, and do so in a timely manner.
In particular, I will discuss why we chose open-source models like Llama3, the merits of frameworks such as LangChain and Vanna, how we ensured we were privacy and risk-compliant, and the infrastructure to build and deploy this.
This session was from our in-person Sandbox Session with 9fin.