Open Source Dapr Adds Large Language Models Interface for Integration with Applications

Diagrid, provider of enterprise tools and services for building secure, reliable, and portable applications, today announced details of the upcoming release of Dapr, a Cloud Native Computing Foundation project maintained by Diagrid, Microsoft, Intel, Alibaba, and others.

Also Read: AI helps Data Engineers be Distinguished Data Engineers

“An additional benefit of Dapr is that it protects developers as AI technology evolves so it’s possible to switch one LLM technology for another with zero impact on the application. As enterprises strive to maintain competitiveness, architecting AI into applications so it can evolve is critical.”

As organizations seek to take advantage of artificial intelligence (AI), they face a critical challenge of how to reliably customize large language models (LLMs) and integrate AI into their applications. The Dapr 1.15 new Conversation Application Programming Interface (API), will enable developers to rapidly build AI-enhanced applications powered by customized LLMs. Using the conversation API reduces the complexity of interacting with LLMs and enables critical security and reliability functions with capabilities like prompt caching and PII data obfuscation. Also, with Dapr 1.15, the Workflow API that was in beta is upgraded to production-ready status. Workflows enable developers to reliably orchestrate microservices – resulting in reliable, long-running, stateful applications.

Also Read: AI in Research: Transforming Processes and Outcomes

“Developers are increasingly being tasked with customizing generic LLMs. Using Dapr, they can code workflows to reliably orchestrate retrieval augmented generation (RAG) pipelines for LLM customization, then use the conversation API to query the LLM, with built-in prompt caching,” said Mark Fussell, co-creator of Dapr and CEO of Diagrid. “An additional benefit of Dapr is that it protects developers as AI technology evolves so it’s possible to switch one LLM technology for another with zero impact on the application. As enterprises strive to maintain competitiveness, architecting AI into applications so it can evolve is critical.”

The CNCF announced Tuesday that the Dapr project achieved graduation from its previous incubation status. Graduation is the highest level a project can attain in CNCF and indicates broad user adoption and maturity. In order to achieve this milestone, Dapr was recognized for its: strong committee governance; security as a fundamental consideration; excellent documentation and contribution guidelines; and an open, engaging community. Tens of thousands of organizations use Dapr to run mission critical applications. Go here to read the blog post with more information.

[To share your insights with us as part of editorial or sponsored content, please write to psen@itechseries.com]