infrastructure setup. It offers GPU access and CPU resources that automatically scale based on workload demand, supporting applications such as large language model inference, model fine-tuning, and high-throughput computation tasks. Modal includes sandbox environments that execute isolated code for testing, evaluation, and agent-based workflows. The platform also integrates observability features that allow monitoring of logs, execution states, and performance metrics across workloads. It supports multi-cloud GPU scheduling and dynamic resource allocation to optimize compute availability across regions. Modal is used by teams building AI applications that require flexible compute scaling, low-latency execution, and simplified deployment workflows.