Overview
Coiled is a cloud computing platform optimized for Python-based data science and engineering. It simplifies scaling workloads by abstracting away the complexities of infrastructure management like Kubernetes and Docker. Users can define and spin up clusters using Python code, automatically distributing their code and environment across cloud machines. Coiled integrates seamlessly with existing Python workflows and tools like Dask, Xarray, and Prefect. Use cases include processing large tabular datasets, training machine learning models, performing geospatial analysis, and running production ETL pipelines. It supports various cloud providers (AWS, GCP, Azure) and offers features like custom networking, private PyPI, spot instances, and integration with custom AMIs. Coiled's architecture emphasizes ephemeral infrastructure, ensuring that users only pay for active compute time.
