SageMaker Serverless Inference using BYOC

As we already know, SageMaker can do basically everything from creating, training, deploying, and optimizing ML models. You can use built-in algorithms and models, browse AWS Marketplace to find specific model packages, or simply create your own - train it using SageMaker and deploy it. Everything is streamlined and organized from start to finish. However, in some circumstances we want a completely custom solution. The idea is to bring our own packages and models i.e.