Modal

About Modal

Modal is a serverless platform designed specifically for AI teams. It allows users to run generative AI models, large-scale batch jobs, job queues, and much more. Users can bring their own code, and Modal takes care of running the infrastructure.

Background and Development

Modal was developed with the aim of simplifying the process of running data-intensive applications. It allows users to fine-tune for any domain, on any data, and supports AI inference, fine-tuning, and batch processing. The platform was built from scratch in Rust for the fastest cold-start times and can scale to hundreds of GPUs and back down to zero in seconds.

Core Features and Capabilities

Modal offers top-of-the-line hardware, including Nvidia A100 and H100 GPUs. It provisions compute instances and installs hardware drivers for users, and also provides cloud storage. Users can mount weights and data in distributed volumes and access them wherever they’re needed. The platform also supports job scheduling, web endpoints, and provides real-time observability.

User Experience

Modal provides an intuitive user interface and a seamless user experience. It allows users to deploy functions to the cloud in seconds, with custom container images and hardware requirements. Users can express container images and hardware specifications entirely in code, and provision network volumes, key-value stores, and queues with ease.

Applications and Use Cases

Modal is used by a wide range of professionals and organizations to power data-intensive applications. It is used by companies like Substack and Ramp for running code on hundreds of GPUs in parallel and for running data-intensive projects respectively.

Impact and Future Outlook

Modal has had a significant impact on its users and the industry, enabling them to run data-intensive applications more efficiently and effectively. The platform continues to evolve, with user feedback playing a crucial role in shaping its future development directions.