Curious Nova

Explore, Learn, Excel – Personalized Education with Gamified Goals and Rewards.

Comment

Curious Nova: Revolutionizing Personalized Learning Through AI

Inspiration

About a year ago, I joined a Master’s in Business Administration program in Nepal, driven by my deep love for learning and a desire to dive deeper into the business world. Having run my own IT business for years, I was excited to expand my understanding of areas like management, economics, and business strategy.

But soon after I joined, I realized that the traditional, rigid classroom style didn’t resonate with me. Most of the classes felt outdated and monotonous. As someone who had never formally studied subjects like economics or spreadsheets, I found it especially hard to stay engaged. The content wasn’t packaged in a way that sparked my curiosity or helped me learn in a way that made sense for me.

So, I decided to take learning into my own hands when I saw the Perplexity hackathon announcement.

Combining my IT experience and my growing fascination with the possibilities of AI, I began building something I wish I had from day one — Curious Nova — a personalized, fun learning tool powered by AI. The idea was simple but powerful: learning should feel like play, not pressure. The tool adapts to each learner’s pace, interests, and style—making complex topics more engaging and digestible.

This journey has been a reminder that education doesn’t have to look one certain way. With the right tools and mindset, we can reshape learning into something exciting, accessible, and deeply personal.

What It Does

Curious Nova is an AI-powered learning platform that creates a unique educational journey for every user. The platform:

  • Analyzes learning preferences through initial assessments and interaction patterns
  • Generates personalized content paths using Perplexity's real-time web search capabilities
  • Delivers adaptive quizzes that adjust difficulty based on performance and understanding gaps
  • Provides contextual AI assistance through an intelligent chat system that answers questions with cited sources
  • Tracks progress visually and gives milestone celebrations
  • Gamifies learning through a comprehensive badge system rewarding completion, consistency, and exploration
  • Ensures content quality by prioritizing scholarly sources and providing verifiable citations

For example, a medical student studying cardiology receives peer-reviewed research papers and case studies, while a hobbyist learning guitar gets beginner-friendly tutorials and practice exercises, all within the same adaptive framework.

How I Built It

Architecture & Design Philosophy

I built Curious Nova with a modular, scalable architecture centered around three core systems:

1. Learner Profile Engine

Users input their age, education, interests, and learning goals. The app uses this data to recommend and tailor content to their level. When starting a topic, users also answer questions about their proficiency and desired depth, allowing the app to design a custom course path.

2. Content Curation System

Using Perplexity's Sonar API, the app fetches real-time, high-quality content. It combines profile data and topic inputs to curate content with Perplexity Sonar Pro, which searches both its model and the internet.

3. Adaptive Assessment Engine

The quiz system dynamically creates questions based on what the user has read. After taking the quiz, the app analyzes their understanding and updates their progress on the topic page.

Technology Stack

  • Frontend: Next.js 14 with TypeScript
  • UI Framework: Tailwind CSS + shadcn/ui
  • Backend: Next.js API routes with Prisma ORM
  • Database: MySQL with optimized indexing
  • AI Integration: Perplexity Sonar Pro API
  • Authentication: NextAuth.js with multi-provider support

Challenges I Ran Into

The Personalization Paradox

Balancing personalization with content discovery was tough. Early versions created “learning bubbles.” I solved this with a curiosity injection algorithm.

API Rate Limiting

Perplexity's rate limits caused content fetch failures. I contacted them and received extra credits to continue development during the hackathon.

Real-time Performance

Generating personalized content in real-time was resource-heavy. I improved performance by:

  • Using background processing for pre-generation
  • Implementing incremental loading with skeleton screens
  • Applying progressive enhancement to load basic content first

Accomplishments That I'm Proud Of

  • Successful MVP Deployment: Built and deployed a working platform in 3 days
  • Real User Validation: 90% of 10 beta testers reported higher engagement
  • AI Integration: 98.5% uptime with fallback systems
  • Inclusive Design: WCAG 2.1 AA standards implemented
  • Performance: Low load times despite dynamic content

What I Learned

Technical Insights

  • API Design: Real-time generation requires smart caching and fallback mechanisms
  • User Experience: Personalization works best when it feels invisible
  • Performance: AI apps must balance responsiveness and intelligence

Product Development

  • User-Centric Design: Direct feedback improved engagement by 40%
  • Iterative Development: Daily demos helped refine the user experience
  • Constraint-Driven Innovation: A 72-hour window focused our priorities

What's Next for Curious Nova

Immediate Roadmap (Next 3 Months)

  • Launch native iOS and Android apps with offline mode
  • Add images and videos for richer content
  • Build learning pattern analytics to recommend study schedules

Medium-term Vision (6–12 Months)

  • Introduce collaborative features like study groups
  • Launch enterprise version for corporate training
  • Enable multilingual learning with AI-powered translation
  • Add voice interface for audio-based learning

Long-term Goals (1–2 Years)

  • Use AR/VR for immersive topic exploration
  • Issue blockchain-based verifiable certificates
  • Collaborate with educational institutions for research
  • Open-source the platform for community-driven content

Impact Goals

Our ultimate vision is to democratize personalized education worldwide. We want to make learning feel like having a personal tutor for every child—fun, tailored, and empowering.

Built With

  • 14
  • ai
  • api
  • api:
  • app
  • bcrypt.js
  • components:
  • css
  • database
  • form
  • hook
  • language:
  • mysql
  • next.js
  • nextauth.js
  • orm:
  • perplexity
  • prisma
  • react
  • router)
  • shadcn/ui
  • styling:
  • tailwind
  • typescript
  • ui
  • vercel
  • zod
Share this project:

Updates