Coding Assistant

Dissertation submitted to
Shri Ramdeobaba College of Engineering & Management, Nagpur
in partial fulfillment of requirement for the award of degree of

Bachelor of Technology (B.Tech)

In

COMPUTER SCIENCE AND ENGINEERING

(Artificial Intelligence & Machine Learning)
By
Priti Dubey (06)
Ansh Agrawal (20)
Harsh Shrivastava (28)
Himanshu Pandey (29)

Guided By
Dr. Preeti Voditel

Department of Computer Science and Engineering

(Artificial Intelligence & Machine Learning)
Shri Ramdeobaba College of Engineering & Management, Nagpur 440 013
(An Autonomous Institute affiliated to Rashtrasant Tukdoji Maharaj Nagpur University Nagpur)
November 2024

Dissertation submitted to
Shri Ramdeobaba College of Engineering & Management, Nagpur

1
SHRI RAMDEOBABA COLLEGE OF ENGINEERING MANAGEMENT, NAGPUR
(An Autonomous Institute affiliated to Rashtrasant Tukdoji Maharaj Nagpur University
Nagpur)

Department of Computer Science and Engineering

(Artificial Intelligence & Machine Learning)

CERTIFICATE

This is to certify that the Thesis on “Coding Assistant” is a Bonafide work of < Priti
Dubey,Ansh Agarwal, Harsh Shrivastava, Himanshu Pandey >, submitted to the
Rashtrasant Tukdoji Maharaj Nagpur University, Nagpur in partial fulfilment of the
award of a Degree of Bachelor of Technology (B.Tech), in Computer Science and
Engineering(Artificial Intelligence & Machine Learning). It has been carried out at the
Department of Computer Science and Engineering, Shri Ramdeobaba College of
Engineering and Management, Nagpur during the academic year 2025-2026.

Date:
Place: Nagpur

Dr.Preeti Voditel Dr. P. Voditel

Name of Guide H.O.D, CSE

Project Guide
CSE

______________
Dr. R. S. Pande
Principal

2
 DECLARATION

We hereby declare that the thesis titled “Coding Assistant” submitted herein, has been
carried out in the Department of Computer Science and Engineering of Shri
Ramdeobaba College of Engineering and Management, Nagpur. The work is original
and has not been submitted earlier as a whole or part for the award of any
degree/diploma at this or any other institution / University.

Date:
Place: Nagpur

Name, roll no and sign of all students

Priti Dubey(06)
Ansh Agrawal(20)
Harsh shrivastava(28)
Himanshu Pandey(29)

3
 APPROVAL SHEET

This report entitled “Coding Assistant” by Priti Dubey,

Ansh Agrawal, Harsh Shrivastava, Himanshu Pandey is
approved for the degree of Bachelor of Technology (B.Tech).

______________________ ______________________
Dr. Preeti Voditel
Project Guide External Examiner

_______________
Dr. Preeti Voditel
H.O.D, CSE

Date:
Place: Nagpur

ACKNOWLEDGEMENT

4
We would like to express our deep and sincere gratitude to our guide Dr. Preeti Voditel
Professor in Computer Science and Engineering(Artificial Intelligence and Machine
Learning) Department, RCOEM, for enabling us to work on this project and offering
valuable advice throughout. Working under his direction was a real privilege and honor.
We are glad for the opportunity to work on this project with him.

We express our sincere gratitude to Dr. Preeti Voditel , Head of the Department of
Computer Science(AIML) Department, RCOEM for his guidance. Talent wins games,
but teamwork and intelligence win championships. We would like to use this
opportunity to express our heartfelt gratitude to everyone who helped us complete this
project.

Priti Dubey -06 Ansh Agrawal-20

Harsh Shrivastava-28 Himanshu Pandey-29

ABSTRACT

5
The objective of this project is to develop an AI-powered Coding Assistant tailored for

interview preparation and real-time code debugging. The platform offers topic-wise

curated coding questions, AI-assisted debugging support, and progress tracking. The

assistant is powered by advanced AI models integrated using Together AI, with real-

time code execution and analysis via Judge0. The frontend is built with Next.js and

Tailwind CSS for a smooth user experience, while MongoDB ensures scalable data

storage and JWT tokens secure user authentication. This system aims to enhance

preparation efficiency for technical interviews and foster better coding practices among

users.

TABLE OF CONTENTS

Content Page No.

6
 Acknowledgements 5
Abstract 6
1.0 Introduction 8
1.1 Background 9
1.2 Motivation 10
1.3 Objectives 11
1.4 Technologies used 12
1.5 Flow of work 14
1.6 Functional Specifications 16
1.7 Scope of the Project 18
1.8 Timeline 20
1.8 Summary 22
1.8 References 23

1.0 Introduction

Modern software development is becoming increasingly complex, with

developers needing to juggle multiple programming languages, frameworks, and
tools while meeting tight deadlines and maintaining high code quality. This
complexity often leads to increased debugging time, steep learning curves for

7
beginners, and reduced overall productivity. To address these challenges, there is
a growing need for intelligent tools that can assist developers throughout the
coding lifecycle—from understanding requirements and writing code to
debugging and documentation.

This project focuses on building an intelligent coding assistant platform designed

to enhance the efficiency and accuracy of software development tasks. By
integrating machine learning, natural language processing (NLP), and static code
analysis, the platform provides features such as real-time code suggestions,
syntax and semantic error detection, code explanations, and natural language
query support. The platform is language-agnostic and supports seamless
integration with common code editors and development environments.

The system architecture facilitates scalable data ingestion (e.g., user queries,
code snippets), efficient model inference, and intuitive user interaction. It
supports advanced models like transformer-based architectures for understanding
code context and intent. By assisting developers in writing, reviewing, and
understanding code more effectively, the solution aims to reduce context-
switching, accelerate development, and serve as a valuable learning companion
for new programmers.

This report outlines the motivations, architecture, key components,

implementation details, and evaluation of the coding assistant platform, along
with its potential applications and future enhancements.

1.1 Background

As the software industry continues to evolve, developers are expected to

manage increasingly complex codebases, adopt new technologies rapidly,
8
and maintain high productivity. Traditional development tools, such as
basic IDEs or static code analyzers, often provide limited assistance and
lack contextual understanding of a developer’s intent. Furthermore,
beginners face significant barriers when learning programming due to the
absence of real-time guidance and intuitive feedback mechanisms.

Recent advancements in artificial intelligence, particularly in machine

learning and natural language processing, have opened new possibilities for
intelligent development support. Leveraging these technologies, it is now
feasible to build systems that understand natural language queries, analyze
code semantics, and provide context-aware suggestions, explanations, and
fixes.

This project addresses the need for such intelligent assistance by

developing a coding assistant platform that acts as a virtual programming
partner. It establishes the backend foundation capable of understanding
user input, interpreting programming context, and delivering useful, timely
responses—whether in the form of code completion, error resolution, or
documentation support. The aim is to reduce the friction in the
development process and empower both novice and experienced
developers with smarter tools.

1.2 Motivation

The motivation behind developing the Coding Assistant stems from

common challenges faced during technical interview preparation and
coding practice. Aspiring developers often struggle with finding structured

9
learning material, identifying and correcting code errors, and understanding
complex programming concepts without proper guidance or feedback.

Manual debugging, searching for solutions online, and a lack of real-time

assistance significantly slow down the learning and development process.
This becomes even more daunting for beginners, who may not yet have the
experience to recognize patterns in errors or optimize their approach to
problem-solving.

With advancements in cloud computing and AI, especially platforms like

Together AI and tools like Judge0, it is now possible to provide real-time,
intelligent feedback on code and interview questions. Our Coding Assistant
is designed to bridge the gap between learning and implementation by
offering:

 Curated topic-wise questions,

 Instant code evaluation,

 AI-powered debugging suggestions, and

 Progress tracking tailored to each user.

By combining these features in a clean, responsive interface using Next.js,

Tailwind CSS, and MongoDB, the platform aims to reduce inefficiencies,
foster continuous improvement, and serve as a smart companion for
developers preparing for interviews or enhancing their coding skills.

1.3 Objectives

• Develop a modular and scalable architecture using Next.js API routes to

support real-time coding assistance.
• Design a clean, responsive user interface using Tailwind CSS for

10
seamless interaction with features like a code editor and debugging
console.
• Integrate Judge0 API for real-time code execution and error detection
across multiple programming languages.
• Leverage Together AI for intelligent code suggestions, explanations, and
optimization based on user input.
• Implement JWT-based authentication for secure user login, session
tracking, and role-based access.
• Enable topic-wise question browsing, code submission, and feedback
collection for continuous learning.
• Use MongoDB to store user data, coding history, and performance
metrics efficiently.
• Ensure extensibility for future features such as leaderboard integration,
user analytics, and multi-language support.
• Provide real-time feedback and progress visualization through a
personalized dashboard.
• Build RESTful endpoints for flexible integration with browser extensions
or IDE plugins in future enhancements.

1.4 Technologies Used

The coding assistant platform integrates a variety of technologies and libraries

spanning natural language processing, machine learning, backend development,
and code analysis. Below are the key technologies used:

1. Next.js
11
 o React-based frontend framework for building fast and scalable web
applications.

o Enables server-side rendering and API routes for backend logic.

2. Tailwind CSS

o Utility-first CSS framework used for rapid and responsive UI

development.

o Offers design consistency and customization through utility

classes.

3. Judge0

o Open-source API for executing and evaluating code in multiple

programming languages.

o Powers the real-time debugging and code execution features of the

assistant.

4. JWT Tokens

o Used for secure authentication and authorization.

o Ensures that user sessions and data access are protected and role-
based.

5. Together AI

o Provides AI capabilities for generating interview questions,

offering explanations, and debugging suggestions.

o Powers the NLP-based reasoning engine behind user query

responses.

6. MongoDB

o NoSQL database used to store user data, question banks, code logs,
and progress tracking.

o Offers scalability and flexible schema design for rapidly evolving

data structures.

7. React Final Form

o Form handling library integrated with the UI to manage validation,

submission, and dynamic inputs.

12
 o Supports real-time feedback and ease of integration with custom
components.

8. Axios

o Promise-based HTTP client for making API calls between frontend

and backend.

o Used to interact with Judge0, backend APIs, and AI services

securely.

1.5 Flow of Work

The flow of work for the store demand forecasting system follows a
structured, modular pipeline—from data collection to model deployment.
Below is a detailed breakdown of each stage:

1. User Query or Code Input

13
  Users interact with a web interface built using Next.js and Tailwind
CSS.

 Input include:

o Code snippets for debugging or explanation

o Problem statements or prompts for code generation

 A syntax-highlighted code editor provides a seamless coding

experience.

2. Preprocessing & Validation

 Submitted code is validated for basic syntax using in-browser logic

before submission.

 Code language and content type are identified and prepared for
processing.

 The query is checked for completeness and context before being sent
to Judge0 or Together AI.

3. Code Execution and AI Response

 Judge0 API handles real-time code execution and returns compiler

output or runtime errors.

 Together AI provides intelligent suggestions including:

o Code explanations

o Debugging hints

o Best-practice alternatives or optimized code

 Best-practice alternatives or optimized code

4. Response Formatting

 Results are structured with proper syntax highlighting, formatting,

and messaging.

 AI responses are shown alongside execution output, enabling side-

by-side learning.

14
 5. User Interaction and Feedback

 Users view results through the dashboard and may rate or comment
on suggestions.

 Follow-up questions or retries can be submitted in the same

interface.

6. Logging & Analytics

 MongoDB stores session history, code attempts, and user activity.

 Interaction logs are used to generate performance graphs and guide

future AI improvements.

7. Backend and API Management

 Next.js API routes serve as backend logic handlers for code

submission, question retrieval, and analytics.

 JWT tokens are used for securing user authentication and session
state.

8. Deployment & Scalability

 The full-stack app is deployed via platforms like Vercel or Render

for seamless CI/CD.

 The architecture supports modular updates (e.g., adding languages or

switching AI providers) without disrupting the system.

1.6 Functional Specifications

1. Query Input Interface

• A responsive and user-friendly web interface built using Next.js

15
 and Tailwind CSS.
• Allows users to enter:
o Topic-wise interview questions
o Code snippets for debugging
• Includes a syntax-highlighted editor for ease of input.

2. Code Submission and Execution

• Integrated with the Judge0 API to compile and execute code in
real-time.
• Supports multiple programming languages for versatile testing.

3. AI-Powered Assistance
• Uses Together AI to generate debugging tips, code explanations,
and alternate solutions.
• Capable of understanding and responding to code and natural
language queries.

4. Dynamic Output Interface

• Displays AI and Judge0 output clearly with syntax formatting.
• Provides readable feedback and suggestions.

5. User Authentication and Session Management

• Utilizes JWT tokens to securely manage user sessions.
• Ensures role-based access for users and admins.

6. Progress Tracking and Reports

• Logs solved questions, debugging attempts, and session data into
MongoDB.
• Displays performance summaries and topic-wise improvement
graphs.

16
 7. RESTful API Endpoints
• Backend routes created with Next.js API for secure handling of
code, user data, and questions.
• API endpoints return JSON responses for frontend integration.

8. Documentation and Help

• Includes clear user documentation and walkthroughs.
• Admin can add questions and manage user roles.

9. Cloud-Ready and Scalable

• Built with modular components deployable on Vercel or Render.
• Supports containerization for scaling across environments.

1.7 Scope of the Project

17
The scope outlines the functionality and limitations of the Coding Assistant
platform during its initial phase and highlights potential areas for future
enhancements.

Current Phase:

 Topic-wise interview question repository covering data structures,

algorithms, databases, and system design.

 Real-time code debugging and execution using the Judge0 API.

 AI-powered assistance using Together AI to suggest fixes, explanations,

and optimized solutions.

 Frontend application developed using Next.js and styled with Tailwind

CSS for a clean, responsive user interface.

 Backend logic handled through Next.js API routes with JWT-based

authentication for user session management.

 MongoDB integration for storing user profiles, code attempts, and

performance metrics.

 Generation of user progress reports including question history and

accuracy trends.

Future Enhancements:

 Integration with popular IDEs like VS Code via extensions for in-editor
debugging.

 Addition of user performance analytics for personalized question

recommendations.

 Gamified coding challenges and leaderboard features to increase user

engagement.

18
  Support for code similarity checks and plagiarism detection.

 Expansion to more languages and frameworks, covering both frontend

and backend technologies.

 Admin dashboard for question management, analytics visualization, and

user moderation.

1.8 Timeline

19
The weekly timeline describes the planned progression of the development and
integration of components for the AI-based coding assistant.

Week 1: Project Planning and Requirement Analysis

 Define the system architecture and target users.

 Identify and finalize the tech stack (Next.js, Judge0, MongoDB, etc.).

 Prepare feature list: interview questions, debugging, progress tracking.

Week 2: UI Design & Frontend Setup

 Set up project using Next.js and Tailwind CSS.

 Design responsive forms and editor components for user inputs.

 Create layout for user dashboard and question navigation.

Week 3: Backend API Development

 Set up API routes for question fetch, user authentication, and code
submissions.

 Integrate MongoDB for data storage.

 Implement JWT-based authentication system.

Week 4: Judge0 and Together AI Integration

 Integrate Judge0 API for code execution and debugging feedback.

 Connect to Together AI for natural language processing and AI

recommendations.

Week 5: Question Bank and Progress Tracker

 Add curated topic-wise interview questions to the database.

 Implement logging of user activity and question attempt history.

 Design progress tracking dashboard and reporting logic.

Week 6: Testing and Bug Fixing

20
  Perform unit and integration testing.

 Validate all API endpoints and user flows.

 Resolve UI/UX bugs and backend errors.

Week 7: Deployment

 Deploy frontend and backend on cloud (e.g., Vercel, Render).

 Ensure database and APIs are properly connected in production.

 Secure environment with proper authentication and API rate limits.

Week 8: Documentation and Final Review

 Write user manuals and internal documentation.

 Create demo video and final presentation.

 Conduct a full system review and freeze development.

21
1.9 Summary

The AI-Based Coding Assistant is a comprehensive platform aimed at supporting

users in preparing for coding interviews and improving programming skills. It
offers curated topic-wise interview questions, AI-powered debugging assistance,
and progress tracking. The system utilizes Together AI for intelligent
suggestions, Judge0 for code execution, and MongoDB for managing user data.
The frontend is developed using Next.js and Tailwind CSS, while JWT tokens
ensure secure access. By integrating these technologies, the platform provides an
efficient, scalable, and user-friendly solution for coding practice and assessment.

22
1.10 References
1. Next.js Documentation. https://nextjs.org/docs
2. Tailwind CSS Documentation. https://tailwindcss.com/docs
3. Judge0 API Documentation. https://judge0.com
4. JWT.io Introduction to JSON Web Tokens. https://jwt.io/introduction
5. Together AI Developer Documentation. https://platform.together.xyz/docs
6. MongoDB Documentation. https://www.mongodb.com/docs
7. React Final Form Documentation. https://final-form.org/react
8. Axios GitHub Repository. https://github.com/axios/axios
9. REST API Design Best Practices. https://restfulapi.net
10. OpenAI GPT Models Overview. https://platform.openai.com/docs/models
11. CodeMirror: The code editor component. https://codemirror.net
12. Git & GitHub Guides. https://docs.github.com

23