# What You Can Do with Kinesis

**Run AI and LLM workloads**\
Deploy, fine-tune, and serve machine learning models using high-performance GPU infrastructure, with the flexibility to scale as demand changes.

**Build and operate data pipelines**\
Execute batch and streaming workloads efficiently, with the ability to adapt compute resources to changing data volumes.

**Optimize cost and utilization**\
Choose between on-demand, usage-based compute or pre-configured resources, and gain a unified view of performance and spend across your workloads.

**Leverage existing infrastructure**\
Connect your own compute resources to Kinesis and manage them alongside platform-provided capacity, all within a single control plane.

**Standardize your runtime environment**\
Use container-based deployments to ensure consistency across development, testing, and production environments.