LakeSail

LakeSail provides a Rust-native, multimodal compute engine at its core, designed to unify batch processing, streaming, and AI workloads. It supports Spark-compatible code via Spark Connect and runs in AWS accounts, with both open-source and managed deployment options. The platform emphasizes open formats, data governance, and a design intended to reduce migration risk while preserving familiar interfaces.

LakeSail is built to run SQL, Python, and AI data pipelines on Sail’s engine, enabling users to execute Spark SQL and DataFrame code through Spark Connect, while keeping access to Iceberg and Delta Lake formats. It offers in-process Python execution and a stateless worker model, aiming to simplify security, scalability, and cost predictability.

The engine is Rust-native and designed to remove JVM bottlenecks, offering instant startup, no garbage collection pauses, and scale-to-zero behavior. LakeSail also provides an agent layer with an MCP server, lakehouse branching, and audit trails as core features, not add-ons.

Deployment options include Open Source Sail, the LakeSail Platform (managed in AWS BYOC), and Enterprise deployments. The platform supports autoscaling, per-second billing, and no minimum spend, with options for on-premises deployment and enterprise SSO where applicable.

Use cases

• Run Spark analytics with SQL and DataFrame code
• Prepare AI datasets and embeddings
• Embeddings and model scoring within pipelines