Modal

Modal is a cloud platform that lets developers run Python functions in serverless GPU containers with a simple decorator-based API. AI models, batch processing jobs, and background tasks defined in Python can be run in the cloud without managing infrastructure, Docker, or Kubernetes. ML engineers, AI application developers, and data scientists use Modal to deploy inference endpoints, run training jobs, process data at scale, and run scheduled AI workloads without DevOps overhead. Models can access GPU resources only when they run, eliminating idle infrastructure costs. Modal's developer-experience focus (deploy a function with a decorator, run it in the cloud) represents a different approach to MLOps: rather than managing model serving infrastructure, developers write standard Python and Modal handles the execution environment, scaling, and GPU allocation automatically.