Situation
Our client, a large financial institution in the Netherlands, is a multinational banking and financial services company that employs more than 17000 FTE worldwide. The client is scaling up its data science capability by increasing the amount of ML (Machine Learning) teams throughout the organisation. They struggle to move beyond the proof of concept phase because of dependencies of the IT department, lacking DevOps engineering skills in the teams and elaborating on repetitive security and compliance processes. Management lacks insights into the ML efforts throughout the bank.
Approach
We helped in realising a centralised platform for the operationalisation of ML models from various teams throughout the bank. ML teams can develop models in their own (federated) environment and easily publish their model to the centralised ML Ops platform where it gets operationalised automatically through end-to-end pipelines (incl. monitoring, logging etc.).
Results
The need for cloud DevOps engineering skills in the data science teams is eliminated, allowing them to work more independently from IT and greatly reducing their time to market. Management is getting insights into the use-cases that are in production on this central platform, and are ensured that their decisions are backed by optimal ML output.
