Kelvin Vector

Kelvin Vector

Many organizations want to leverage machine learning and artificial intelligence to improve their business processes. However, the process of building and deploying real workflows is often undermined by the difficulty of managing data at scale. Vector databases are a new class of data store specifically designed to help solve this problem.

Kelvin Vector is a library that provides a vector storage and calculations for the Kelvin Legal Data OS. It is designed to be simple, powerful, and seamless to integrate into existing workflows. Kelvin Vector can be used as:

  • A high-performance transient or durable store for Python applications
  • A high-performance transient or durable store via API
  • Through integrated vector search in Kelvin Document Index

The following durable storage options are available:

  • Filesystem
  • Postgres
  • SQL Server

Use Cases

Kelvin Vector is designed to simplify the following common use cases:

  • Document Search
  • Document or Text Clustering
  • Document or Text Classification
  • Feature Engineering
  • Feature Storage

Library and Framework Integrations

Kelvin Vector is designed to integrate not just with the Kelvin Legal Data OS, but also with other common data science and machine learning libraries or frameworks, including the following:

  • pandas
  • scikit-learn
  • PyTorch
  • TensorFlow
  • transformers (Hugging Face)
  • Apache Arrow
  • Apache Spark (Databricks)
  • Dask

API and Library Documentation

You can learn more about Kelvin Vector by reviewing the PyDoc documentation and OpenAPI schemas provided with the library. To access the PyDoc documentation, you can run the following command:

kelvin --docs kelvin.vector