Arcus Documentation
Back to homepage

Introduction to Arcus


The Arcus Data Platform is a turnkey model improvement solution that seamlessly improves your AI applications’ performance by discovering and integrating relevant features within your internal data sources, valuable external data and synthetic data that transparently improves your ML Model’s performance. Your AI applications are only as good as the data they see and by providing your applications with more high-value real world data and signals, you improve their ability to accurately model and understand their task at hand. Arcus provides multiple offerings for different ways in which you can interact with the Data Platform to improve your AI applications.

Exchange Overview

The Arcus Data Platform automatically matches AI applications with data that improves their performance. By connecting to the Data Platform, your AI applications are automatically matched to and joined with relevant datasets, which are then empirically validated through the trialing process. Arcus’ client libraries develop an understanding of your application and first-party data to predict the highest value external data to consume and then run experiments to determine the quantifiable impact that data has against your task. Based on the actual results produced, you can choose what data to consume and integrate directly into your application.

Model Enrichment

Using Arcus Model Enrichment, you can seamlessly improve the performance of your AI models. Models are automatically enriched with high value external signals and data to help them better model real world applications. Using the Arcus Model SDK, the data you consume from the platform is automatically integrated into your existing ML workflows for training and inference, all while preserving the privacy of your data.

Prompt Enrichment

Using Arcus Prompt Enrichment, you can seamlessly improve the performance of your Generative AI applications. The prompts for your Generative AI applications are automatically enriched through the process of Data Augmented Generation with signals and data from the platform, which are provided to the model as prompt context.

Next Steps

The Arcus Client Libraries allow you to consume data from the Arcus Data Platform without any changes to your core ML workflows. Your models and data stay where they are, while the client libraries integrate popular ML frameworks and libraries with the Data Platform to transparently pull in high value data and signals and simply improve your AI performance.

Explore the key concepts in the Arcus platform to learn more.