> ## Documentation Index
> Fetch the complete documentation index at: https://docs.lancedb.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Multimodal Feature Engineering with Geneva

> Learn how to do multimodal feature engineering in LanceDB Enterprise to transform raw data into meaningful features for AI models.

<Badge color="red">Enterprise-only</Badge>

When working with multimodal data at scale, [LanceDB Enterprise](/enterprise) makes it easy
to define, extract, and transform raw data into useful information and features for your
AI applications. LanceDB Enterprise's *Multimodal Feature Engineering* package is designed to improve
the productivity of AI engineers operating at immense scale.

With an API designed to leverage LanceDB's optimized data storage and retrieval, it
streamlines prototyping extraction and transformation tasks, performing experiments, exploring your
data, scaling up execution, and moving to production.

<Card>
  Feature Engineering and the `geneva` Python package are currently only available as part of
  [LanceDB Enterprise](/enterprise). Please [contact us](mailto:contact@lancedb.com) if you're interested
  in scaling up your feature engineering workloads for your AI and multimodal use cases.
</Card>

The `geneva` package uses Python [User Defined Functions (UDFs)](/geneva/udfs/udfs) to define features
as columns in a Lance dataset. Adding a feature is straightforward:

<Steps>
  <Step>
    Prototype your Python function in your favorite environment.
  </Step>

  <Step>
    Wrap the function with a small UDF decorator (see [UDFs](/geneva/udfs/udfs)).
  </Step>

  <Step>
    Register the UDF as a virtual column using `Table.add_columns()`.
  </Step>

  <Step>
    (Optional) Configure where the UDF will run: locally, on a Ray cluster, or on a Kubernetes cluster with KubeRay (see [Contexts](/geneva/jobs/contexts)).
  </Step>

  <Step>
    Trigger a `backfill` operation (see [Backfilling](/geneva/jobs/backfilling/)).
  </Step>
</Steps>

<Tip>
  You can build your Python feature generator function in an IDE or a notebook using your project's Python versions and dependencies. `geneva` will automate much of the dependency and version management needed to move from prototype to scale and production.
</Tip>

## Continue learning

Visit the following pages to learn more about featuring engineering in LanceDB Enterprise:

* **Overview**: [What is Feature Engineering?](/geneva/overview/) · [End-to-end example](/geneva/end-to-end)
* **UDFs**: [Using UDFs](/geneva/udfs/udfs) · [Blob helpers](/geneva/udfs/blobs/) · [Error handling](/geneva/udfs/error_handling) · [Advanced configuration](/geneva/udfs/advanced-configuration)
* **Jobs**: [Backfilling](/geneva/jobs/backfilling/) · [Startup optimizations](/geneva/jobs/startup/) · [Materialized views](/geneva/jobs/materialized-views/) · [Execution contexts](/geneva/jobs/contexts/) · [Geneva console](/geneva/jobs/console) · [Performance](/geneva/jobs/performance/)
* **Deployment**: [Deployment overview](/geneva/deployment/) · [Helm deployment](/geneva/deployment/helm/) · [Troubleshooting](/geneva/deployment/troubleshooting/)

## API Reference

* [`geneva.connect()`](https://lancedb.github.io/geneva/api/) — connect to a Geneva database
* [Connection](https://lancedb.github.io/geneva/api/connection/) — manage tables, views, jobs, clusters, and manifests
* [Table](https://lancedb.github.io/geneva/api/table/) — add columns, backfill, search, and manage table data
* [UDF](https://lancedb.github.io/geneva/api/udf/) — define user-defined functions for feature computation