Skip to main content

Developing

We built the Modular Platform from the ground up to simplify AI development for production and get the most out of your GPUs.

The MAX Python API lets you work at the level of abstraction that best fits your needs. Whether you want to serve and deploy models or build custom neural networks from scratch, MAX offers the tools to get there.

At the top of the stack, MAX gives you immediate access to optimized, production-ready models for common AI tasks like text generation and embeddings.

When you need more control over your model architecture, the neural network module provides high-level building blocks that compile down to optimized computational graphs.

For complete customization, the Graph API lets you define computation at the operation level, while the Driver API manages how that computation executes across different hardware.

Finally, the engine ties it all together by providing the runtime that executes your models with maximum performance.

Guides

Tutorials