Radiant is the Enterprise AI platform to take your idea from prototype to production deployment. From security and governance to scaling and anomaly detection, we make it simple to make AI a critical part of your business. 

Apologies for the delay in our updates, here are the things we’ve delivered in the past two weeks. Our product team released new tools to test and evaluate models and prompt changes. Our new comparison feature makes it easy to compare datasets, using evaluation metrics to determine if your changes impact performance on your application. We’re excited to show your our latest features now available in Radiant.

Product Updates

Dataset and Metric Comparisons

One of the major challenges in making changes to models or prompts to production Generative AI systems is understanding how these changes impact performance. Creating a dataset of key prompt and response pairs in Radiant allows users to measure and score a group of prompt-response pairs and then set up a comparison with datasets to compare metrics and differences in responses. User can create datasets with specific examples from a particular model or set of responses and compare metrics across the datasets.

Use Radiant to provide positive examples with your prompts

Radiant datasets help users collect examples of model behavior that can be used to improve production workflows. Now users can leverage the Radiant API to provide examples directly from Radiant. You can provide models example prompts directly from the platform making it easier to curate and prompt good examples.

Streamlined creation of datasets and detailed trace tagging

Users can now select multiple events from their logs and easily add or remove them to a dataset. Users can simply click the “new selection” button to add log events to their dataset. Each log trace now also contains more information about the type of event including chat completion or embeddings c

Restricted log view gets an improvement

For projects marked for sensitive data we made it easier to hide and reveal data. By default sensitive projects will have a limited preview button that hides log details. Users with the correct permissions can reveal the full log details by clicking the button.

New Blog Post: The best use cases for AI are where there's asymmetric risk

In this post we explore how to plan successful Generative AI projects and the kinds of use cases that make for better business outcomes: when the risk is low and the reward is high.

About Radiant

Radiant is the Enterprise AI platform to take you from idea to production deployment. From security and governance to scaling and anomaly detection, we make it simple to make AI a critical part of your business. 

Try out a demo here, sign up here to get your own instance, or reach out to our founders directly at [email protected].

We’re also hiring. If you know of someone great that is interested in helping every company build AI into their products and operations, we'd love an introduction.