DSA Mentor

DSA Mentor is an AI-powered Chrome extension that analyzes LeetCode solutions with Gemini, tracks progress via viaSocket, and sends personalized weekly reports to improve DSA skills.

Comment

DSA Review AI – Your Personal Coding Interview Mentor

Inspiration

Every day, thousands of students solve Data Structures and Algorithms (DSA) problems on platforms like LeetCode. However, most only know whether their solution is accepted—they rarely receive detailed feedback on why their approach works, whether it is optimal, or how an interviewer would evaluate it.

As a student preparing for software engineering interviews, I wanted an AI-powered reviewer that behaves like an experienced interviewer rather than just a compiler. This inspired me to build DSA Review AI, a Leetcode extension that instantly reviews selected DSA solutions using the Gemini API.

What it does

The extension allows users to simply select their DSA solution and receive an in-depth AI review within seconds.

It analyzes:

  • Time Complexity
  • Space Complexity
  • Code Readability
  • Logic & Flow Explanation
  • Dry Run of the Algorithm
  • Edge Cases
  • Potential Bugs
  • Better or Optimal Solutions
  • Interview-style Feedback
  • Overall Performance Rating

Instead of just telling users whether the solution works, the extension helps them understand *how to improve as problem solvers.

How we built it

The project is built using TypeScript.

Workflow:

  1. The user selects a DSA solution.
  2. The extension extracts the selected code.
  3. A carefully engineered prompt is sent to the Gemini API.
  4. Gemini performs a detailed code review and returns a structured JSON response.
  5. The extension renders the review.

Using structured JSON instead of plain text made the responses consistent and easier to display.

Gemini API Integration

Gemini serves as the intelligent reviewer behind the project.

Rather than asking Gemini to simply explain code, we designed prompts that instruct it to evaluate solutions like a technical interviewer by analyzing complexity, algorithmic patterns, optimization opportunities, readability, and correctness.

Prompt engineering played a major role in improving response quality and consistency.

Challenges

Some of the biggest challenges included:

  • Designing prompts that produced reliable and structured responses.
  • Making AI feedback consistent across different coding styles.
  • Rendering detailed reviews in a clean and readable VS Code interface.
  • Balancing comprehensive analysis with fast response times.

What we learned

During this project we learned:

  • Extension Development
  • Gemini API Integration
  • Prompt Engineering
  • Structured JSON generation from LLMs
  • Building developer tools powered by AI
  • Designing user experiences for programming workflows

Future Improvements

We plan to extend the project with several advanced features:

  • Support for multiple programming languages
  • Inline code annotations
  • Visual algorithm flow diagrams
  • Weekly coding progress reports
  • Personalized improvement recommendations
  • One-click comparison with optimal solutions
  • AI-generated practice problems based on weaknesses
  • Integration with LeetCode and Codeforces profiles

Impact

DSA Review AI transforms accepted code into a learning opportunity. Instead of simply solving problems, developers receive detailed feedback similar to what they would get from an experienced interviewer, helping them prepare more effectively for coding interviews while continuously improving their programming skills.

Built With

Share this project:

Updates