We primarily use it for real-time use cases where we have used Lambda with the API Gateway. So, the backend is a Lambda server. We are implementing a serverless architecture with the API Gateway.
We've implemented risk score models within the Lambda. Thus, whenever someone requires a risk assessment for a prospect in need of a loan, they access our Amazon API Gateway. In the backend, it triggers the Lambda service.
So, within the Lambda, our ML model predicts whether the customer falls into the high-risk or low-risk category. Then, it sends a response to our frontend application, which is an Android-based application.
Currently, our use cases are relatively simple. Even though we host ten different models, the use cases are similar. We have configured the API gateway for various URLs and receive hits from different services like tractors, two-wheelers, and mobiles.
The most valuable aspects are its user-friendliness and simplicity. These two features make it easy to work with.
Additionally, it's cost-effective, even as our workload and the number of APIs we manage have increased.
The integration with other API services, such as Lambda, EC2, and Kubernetes, is also incredibly useful.