CodeSense

Inspiration

As a student learning to code, I often found it hard to identify mistakes or improve my coding style without constant feedback. This inspired me to build CodeSense, an AI tool that helps students and beginners get instant code reviews, learn better practices, and improve their coding confidence.

What it does

CodeSense reviews your code automatically and provides: Error Detection: Finds possible bugs or syntax mistakes. Suggestions: Gives feedback on how to write cleaner and more efficient code. Learning Tips: Explains what each mistake means and how to fix it. Code Score: Rates your code quality to help you improve over time.

How we built it

I built CodeSense using: Python for backend logic and AI model integration Streamlit for an easy-to-use web interface Flask to connect frontend and backend CodeBERT and GPT models for code understanding Static analysis tools like Pylint to check for basic errors

Challenges we ran into

Making the AI give useful and beginner-friendly explanations Handling different code styles and formats Combining AI suggestions with static analysis results Keeping the app simple and lightweight for fast feedback

Accomplishments that we're proud of

Built an AI tool that can actually review and explain code Created a project that helps students learn coding more effectively Learned how to connect machine learning with real-world applications

What we learned

How AI can be applied to code analysis and education The basics of using NLP models like CodeBERT How to build and deploy interactive AI-based web apps How to write clearer and more maintainable code

What's next for CodeSense

Add support for more programming languages Create a VS Code plugin for real-time AI suggestions Improve feedback with visual explanations and learning tips Make CodeSense public so other students can learn from it too

Built With

• codebert
• css
• flask
• github
• gpt-api
• html
• pylint
• python
• scikit-learn
• streamlit