Inventor(s)

Hem Dutt

Abstract

Software development kits (SDKs) and frameworks are critical to software development as they provide a collection of pre-written code, libraries, and tools that allow developers to create applications more efficiently. Although Large Language Models (LLMs), when combined with Retrieval-Augmented Generation (RAG) techniques, have shown remarkable potential in code generation, applying LLMs and RAG to generate code for SDKs and frameworks remains challenging. For example, computer code is inherently structured and often relies on precise relationships between various elements. Misinterpretation or loss of these relationships can lead to incomplete or incorrect code generation. SDKs and frameworks compound this challenge by introducing additional layers of complex dependencies. To address these challenges, the techniques proposed herein leverage an approach of combining LLMs and RAG to generate code using an artificial intelligence (AI) Coder specifically designed for SDKs and frameworks.

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.

Share

COinS