AI-Enhanced Code Review Assistant for Web Developers : Designing, implementing, and evaluating an AI-powered web application for automated code feedback
Hiththatiyage, Hasitha; Kodikara, Nipuni (2026)
2026
All rights reserved. This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-202603043727
https://urn.fi/URN:NBN:fi:amk-202603043727
Tiivistelmä
The thesis was aimed at developing a web application using artificial intelligence to assist web developers in the code reviewing process. The main aim of the thesis was to design and develop an artificial intelligence code reviewer to help identify coding errors and inefficiencies while providing the developer with the solution in a natural and understandable manner.
Although human code reviews are an essential part of the development of software applications, it could prove to be time-consuming and highly subject to the knowledge of the reviewing person. As of now, applications for instruction and development that could interpret source code and give understandable feedback do not exist. Therefore, the primary objectives of this research paper would be to: (1) identify appropriate AI technology and development platforms for the assessment of the code; and (2) develop an efficient web application that combines user-friendly feedback representation and analysis by AI technology for reviewing.
A key aspect incorporated within the current thesis includes the system architectural design and the choice of appropriate frameworks and technology. The architectural design chosen for the thesis focuses on the client–server model and includes the frontend implementation using React, the backend development using Node.js Express, and the PostgreSQL-based data storage layer managed through the Supabase platform.In the context of the analytics part, the role of large language models (LLMs) accessed via the Groq Cloud API is acknowledged.
Current efforts have been placed on analysing requirements and performing the initial implementation of the system's core design elements. In the future work session, the AI model will then be integrated in its entirety, while evaluation measures for accuracy and usability will also be determined to conduct user testing with both web developers and students.
Although human code reviews are an essential part of the development of software applications, it could prove to be time-consuming and highly subject to the knowledge of the reviewing person. As of now, applications for instruction and development that could interpret source code and give understandable feedback do not exist. Therefore, the primary objectives of this research paper would be to: (1) identify appropriate AI technology and development platforms for the assessment of the code; and (2) develop an efficient web application that combines user-friendly feedback representation and analysis by AI technology for reviewing.
A key aspect incorporated within the current thesis includes the system architectural design and the choice of appropriate frameworks and technology. The architectural design chosen for the thesis focuses on the client–server model and includes the frontend implementation using React, the backend development using Node.js Express, and the PostgreSQL-based data storage layer managed through the Supabase platform.In the context of the analytics part, the role of large language models (LLMs) accessed via the Groq Cloud API is acknowledged.
Current efforts have been placed on analysing requirements and performing the initial implementation of the system's core design elements. In the future work session, the AI model will then be integrated in its entirety, while evaluation measures for accuracy and usability will also be determined to conduct user testing with both web developers and students.