On Continuous Integration / Continuous Delivery for Automated Deployment of Machine Learning Models using MLOps
Satvik Garg, Pradyumn Pundir, Geetanjali Rathee, P. K. Gupta, Somya Garg, Saransh Ahlawat
Abstract
In recent years, model deployment in machine learning is observed to be an interesting area of study. It can be seen as a process similar to the one established for traditional software development. Development and operations (DevOps) incorporating Continuous Integration and Continuous Delivery (CI/CD) have demonstrated to smooth out software advancement and speed up organizations. Nonetheless, employing CI/CD pipelines in an application that incorporates components of Machine Learning Operations (MLOps) has challenging issues, and pioneers in the field settle them with the utilization of exclusive tooling, frequently presented by cloud suppliers. This study gives a higher perspective on the machine learning lifecycle and the vital differences between DevOps and MLOps. We talk about tools and techniques to execute the CI/CD pipeline of machine learning frameworks in the MLOps approach. Subsequently, we deep dive into push and pull-based deployments in Github Operations (GitOps). Open exploration challenges are additionally distinguished and added that can direct future research.