Modal

Modal is a serverless cloud platform that lets developers run Python code, AI model inference, fine-tuning jobs and batch pipelines on GPUs without configuring servers or containers. You annotate a function, deploy it in seconds, and Modal handles autoscaling from zero to hundreds of GPUs while billing per second of compute. Machine learning teams use it for model endpoints, data processing, scheduled jobs and web endpoints, making it a practical alternative to managing Kubernetes or reserving expensive GPU instances.