Inspiration
I was struggling deeply with DSA. The questions didn’t make sense, other people’s solutions made even less sense, and I had no idea how to improve. The advice was always the same: “solve practice sheets,” “watch YouTube explanations,” “XYZ is the best channel for this.”
What I never got was anything personalized.
Ironically, AI, which I expected to help, made things worse. My problem-solving skills started to deteriorate because I stopped thinking. I wasn’t learning anymore, I was chasing solutions instead of embracing the struggle.
That’s when I realized I didn’t just want to build something that “helps with DSA.” I wanted to build something that helps people the right way.
A way where you’re not copy-pasting answers or blindly trusting AI. A way where guidance adapts to you, not the other way around. A way where motivation exists even when progress is slow, especially when it’s slow.
What it does
AlgoDash is a personalized dashboard for everything related to DSA and competitive programming.
Users can sign in and connect their profiles from LeetCode, CodeChef, and Codeforces (at least one required). AlgoDash then visualizes their journey through rating graphs, stats cards, and topic/tag-wise distributions, giving users a clear sense of where they stand, whether they need motivation or a quick reality check.
Beyond visualization, AlgoDash actively guides improvement. It recommends problems and contests tailored to your weak areas. Struggling with greedy problems? Good. AlgoDash will push you toward more of them, because that’s how growth actually happens.
To add another layer of personalization, AlgoDash generates daily AI-powered feedback. This feedback helps users:
- Reflect on previously failed submissions
- Rethink problem-solving approaches
- Discover relevant learning resources
- Set concrete, time-bound goals, broken into quarters and deadlines
Not vague advice like “solve a sheet,” but actionable direction.
However, competitive programmers often face a dangerous trade-off: use AI to save time or preserve real problem-solving ability.
AlgoDash refuses to treat this as a trade-off.
That’s why AlgoDash’s chatbot will never reveal solutions or code. The implementation is yours. The logic is yours.
But if you’re confused about an underlying concept, stuck in reasoning, or unsure how to think and that’s where the chatbot steps in.
Guidance without shortcuts. Help without harm.
Overall, AlgoDash aims to deliver effortless, personalized guidance without paywalls, strict policies, or compromising the learning process.
How I built it
The backend is built using Flask. AI features are powered by Groq, chosen for its fast inference, affordability, and support for customizable response schemas.
User data is fetched using multiple platform APIs (linked below).
Authentication and database management are handled via Supabase, providing reliable cloud storage and a smooth developer workflow.
On the frontend, Jinja is used extensively to dynamically render user-specific data, which is essential for a dashboard-centric application like AlgoDash.
Challenges I ran into
Row Level Security (RLS) issues
While updating feedback timestamps and content, Supabase blocked requests due to RLS violations. I solved this by creating two Supabase clients:
- A public client using the anon key
- A secure client using the service-role key
This allowed safe inserts and updates without compromising security.
Slow loading times
Most pages depended on multiple utility functions, causing noticeable delays before users could even see their stats. I resolved this by implementing simple caching in Flask, significantly improving load times.
Restrictive CodeChef API
CodeChef’s strict bot-crawling policies limited access to detailed user data. To compensate, AlgoDash generates feedback using a combined performance overview, ensuring meaningful insights regardless of platform limitations.
Accomplishments that I’m proud of
Building under a tight deadline I joined the hackathon late and had roughly a week to build, document, and present the project, while balancing academics, and still competed among 1200+ participants.
Creating tech that genuinely helps people AlgoDash was built from personal frustration and shared experiences with friends and online communities. Building something that could meaningfully improve how people learn DSA was deeply fulfilling.
What I learned
Quality beats quantity Instead of adding dozens of disconnected features, I focused on tools that share the same data foundation and reinforce each other, like the dashboard, problem recommendations, AI feedback, and chatbot all work together.
Make it work before making it perfect With limited time, I prioritized functionality and usability over perfection. I focused on features that deliver, tools that guide, and a UI that actually helps users navigate.
Practical caching in Flask This was my first real experience implementing caching, and applying it to a production-like project helped my understanding far better than tutorials ever could.
What’s next for AlgoDash
Public deployment After further testing and iteration, AlgoDash will be made publicly accessible. It was built for users so they should be able to use it.
More platform integrations Planned integrations include HackerRank and Codewars, making AlgoDash useful regardless of a user’s preferred competitive programming ecosystem.
Log in or sign up for Devpost to join the conversation.