Abstract:
QUBIT (Quality Unit-Based Intelligent Testing) utilizes the use of Large Language Models
(LLMs) to boost-up the potential of not only developers but all coding enthusiasts. Its primary focus
revolves around Code Review, an integral step in software development. This project comprises of
two distinct parts. The first component is an extension specifically designed for Visual Code Studio.
It seamlessly integrates with the user's development environment, offering real-time Code Review,
Code Rating, Bug detection, and Code Refactoring. Moreover, it also generates test cases for robust
development along with Code Review. It allows the selection of multiple files for the sake of better
efficiency. The second component of QUBIT manifests as a website, also utilizing the prowess of
LLMs, specifically Gemini, to conduct features involving, Code Review Code Rating, Bug
detection, Code Refactoring. The website not just mirrors the capabilities of the VS Code extension
like a chatbot acting as a virtual assistant which offers dynamic suggestions, it also introduces
additional features. It includes a user-friendly dashboard; QUBIT community and customizable
profiles enhance the overall user experience. Prompt engineering serves as the backbone,
facilitating seamless interactions with the LLMS. Gemini is utilized for all the functionalities other
than test case generation, which harnesses the use of Llama2. QUBIT not only increases the
efficiency of developers but also explores the frontier of AI-driven coding assistance. Both
components of this project aspire to redefine the landscape of coding tools.
