End-to-End ML Platforms

It’s our belief that effective platforms are key to delivering ML and AI at scale. These platforms support data science and ML engineering teams by allowing them to innovate more quickly and consistently. So what is a machine learning platform? A machine learning platform is a set of tools and technologies (backed by a set of practices and processes) established by an organization to support and automate various aspects of the machine learning workflow, including data acquisition, feature and experiment management, and model development, deployment, and monitoring.

Machine learning platforms come in a wide variety of forms. Until recently, they have primarily been found at large technology companies, which have developed their platforms internally, out of necessity, to support increasingly significant investments in machine learning. As the importance of machine learning has become clear to a broader array of enterprises, new commercial and open source ML platform technologies have become available to reduce the barriers to adoption and make the benefits of ML models more accessible.

Some of the benefits of an end to end platform are really also the benefits of MLOps generally, namely:

• Team collaboration and productivity: your team can work together more productively across the organization and across roles;
• Repeatability: If something is working it can be replicated easily.
• Auditability: if everything is tracked, then it can also be audited to ensure good governance;
• Higher quality: The ability to put in policies that address model bias and track quality and accuracy over time can lead to higher quality model outputs or point out degradations so that they can be addressed more quickly.

By abstracting across the experiences of many early platform builders and users, and identifying the capabilities that we see frequently recurring in the platforms that they’ve built, we have identified a core set of platform capabilities.

It’s important to note that these End-to-End ML Platforms encompass three other sub-categories:

• Data Acquisition and Preparation;
• Model Development and Training;
• Model Deployment and Operations.

In addition to the above, they generally include system-wide functions such as:

• ML Infrastructure Orchestration;
• Hardware Accelerator Support;
• Kubernetes Integration;
• Teamwork and Collaboration;
• Enterprise-grade Security;
• Model Transparency (bias and fairness)
• and Governance.