# fastkernels

## Docs

- [Architecture](https://snowflake-84d72a0d.mintlify.app/architecture.md): The L1–L4 operator hierarchy, interface mirroring, and how to extend fastkernels.
- [Benchmarking](https://snowflake-84d72a0d.mintlify.app/benchmarking.md): Three-tier benchmarking infrastructure for kernel development.
- [Building an Agent](https://snowflake-84d72a0d.mintlify.app/building-an-agent.md): Use an LLM to automatically generate and benchmark replacement kernels.
- [Developer Guide](https://snowflake-84d72a0d.mintlify.app/developer-guide.md): Testing against vLLM and adding new model architectures.
- [Experiment Tracking](https://snowflake-84d72a0d.mintlify.app/experiment-tracking.md): Track kernel generations, benchmark results, and agent runs with MLflow.
- [Intro to KB Nano](https://snowflake-84d72a0d.mintlify.app/intro.md)

## OpenAPI Specs

- [openapi](https://snowflake-84d72a0d.mintlify.app/api-reference/openapi.json)