DSA-Whisper

Inspiration

As beginners, solving Data Structures & Algorithms problems often feels intimidating and isolating. Many students quit early because they don’t understand how to think, not because they can’t code. I wanted to build a tool that acts like a mentor — explaining logic step-by-step, breaking down problems, and teaching DSA in a way that feels conversational and human. DSA-Whisper was born from that motivation: to make algorithmic thinking accessible for everyone, especially students who don’t have mentors.

What it does

DSA-Whisper is an AI-powered learning assistant that: Explains DSA concepts in simple, structured steps Solves LeetCode-style problems with reasoning breakdown Generates custom study plans based on the user’s skill level Provides multiple-approach solutions (brute force → optimal) Visualizes algorithms (e.g., pointers, stack flow, recursion tree) Supports debugging: users paste code, and the model explains the error Offers problem-specific hints without revealing the full solution immediately It acts as a “DSA tutor” that adapts to the learner’s pace.

How we built it

Built a Streamlit web app interface for clean interaction Integrated Groq LLM for fast, cost-efficient reasoning on DSA problems Implemented custom prompt-engineering to force: step-by-step explanations multiple solution paths examples for edge case Added a dynamic UI with: topic selector (Arrays, Linked Lists, Stacks, Trees, DP…) structured output boxes (Concept → Approach → Code → Complexity) Used Python modules for small animations (pointer movement, recursion tracing) Deployed on Hugging Face for public access

Challenges we ran into

Getting the LLM to produce consistent step-wise explanations without hallucinations Handling very long problem inputs in Streamlit Prompt failures where the model skipped brute-force or complexity analysis Ensuring code formatting and indentation stayed correc Designing a beginner-friendly UI while still supporting advanced users Time-management during the hackathon with many features to implement

Accomplishments that we're proud of

Built a complete working app that genuinely helps users learn DSA logically Achieved stable, multi-approach answers for almost all major problem patterns Created a smooth, clean and professional Streamlit UI Reduced model hallucination rate through rigorous prompt-engineering Successfully deployed a fully functioning DSA-teaching assistant publicly Received positive feedback from students who tested it

What we learned

How to engineer reliable prompts for algorithmic reasoning How to design educational tools around user learning psychology How to break down complex problems so beginners can understand them UI/UX techniques to present long explanations neatly Practical experience deploying apps to Hugging Face Importance of time-boxing and prioritizing core features during hackathons

What's next for DSA-Whisper

Add code execution sandbox for real test cases Add visual diagram generator for trees, graphs, DP tables Add user progress tracking and weekly practice goals Include interactive quizzes and challenge modes Introduce multilingual support (Urdu, Hindi) Build a personalized roadmap generator using user history Expand into system design explanations

Built With

• groqapi
• huggingface
• python
• streamlit