SkillCraft | Learn. Play. Level Up

Home Page | Awesome UI with smooth animations
leaderboard with XP points
Avatar Badges for Gamification
Time Spent badges with streaks
Lecture Videos with ML monitoring
Courses Status and Suggestions

Inspiration

We saw how traditional education often leaves students disengaged and unprepared for real-world challenges. SkillCraft was born from the idea of fusing game mechanics with AI-driven simulations to create an educational playground where learners level up both skills and confidence.

What it does

Freelancing Simulator 🏗️
- Bid on dummy projects, compete in AI-judged bids, and complete simulated gigs with a virtual client.
Job Simulation Hub 💼
- Apply to mock job postings, face AI-powered GD rounds and interviews, then collaborate on team projects.
Adaptive Learning Path
- Dynamic leaderboards, skill badges, and 24/7 tutor bot keep motivation high

Live Demo: Visit SkillCraft

How we built it

AI/ML Core: Python • TensorFlow • PyTorch • RLlib for adaptive difficulty
Generative AI: GPT-4, Claude 3, Llama 3
Backend: ExpressJs with Supabase for PostgreS services,
Frontend: React with Three.js for 3D experiences

Challenges we ran into

Integrating real-time AI avatars for interviews without noticeable lag
Balancing game mechanics with genuine learning outcomes
Designing a scalable bid-ranking model combining text clarity and numerical regression

Accomplishments that we're proud of

96% accuracy in AI-driven interview coaching with Unity avatars
40% faster skill acquisition through reinforcement learning–powered adaptive lessons
Real-time leaderboards powered by Redis sorted sets, updated in <2 seconds
Personalized topic recommendations using BKT + collaborative filtering

What we learned

Deep insights into student behavior enabled us to craft meaningful learning styles via K-Means clustering
The importance of combining quantitative and qualitative feedback for bid-ranking predictions
Techniques for low-latency AI inference in browser-based simulations