An intelligent AI-powered study companion built with Flask, ML, Deep Learning, NLP, and deployed to production.
Status: ✅ Live & Running | 📊 Full-Stack | 🚀 Production Ready
A complete AI-powered study assistant web application demonstrating all five domains of the AI course curriculum: Python · ML · Deep Learning · NLP · Web Deployment
AI_Study_Pal/
│
├── app.py ← Flask web server (all routes)
├── ai_engine.py ← Complete AI / ML brain
├── requirements.txt ← Python dependencies
├── Procfile ← Render.com / Gunicorn start command
├── render.yaml ← Render.com auto-deploy blueprint
├── runtime.txt ← Python version pin
├── .gitignore
├── README.md
│
├── templates/
│ └── index.html ← Single-page frontend (Jinja2)
│
├── static/
│ ├── style.css ← Dark Academic theme
│ └── script.js ← All frontend JS / API calls
│
└── models/ ← Auto-created; trained models saved here
├── quiz_vec.pkl
├── quiz_clf.pkl
├── topic_km.pkl
├── topic_vec.pkl
├── summarizer.keras
└── feedback.keras
| Feature | Technology | Model |
|---|---|---|
| Study Planner | Pandas date ranges | Rule-based scheduling |
| Quiz Generator | scikit-learn | TF-IDF + Logistic Regression |
| Topic Clustering | scikit-learn | K-means |
| Text Summarizer | Keras Dense NN | Sentence importance scoring |
| Motivational Feedback | Keras Embedding NN | Score-bucket classifier |
| Study Tips | NLTK | Tokenisation + keyword extraction |
| Visualisations | Matplotlib | Pie & bar charts → base64 |
| Data Handling | Pandas | EDA, scheduling, CSV export |
cd AI_Study_Pal# Windows
python -m venv venv
venv\Scripts\activate
# Mac / Linux
python3 -m venv venv
source venv/bin/activatepip install -r requirements.txt⏳ First install takes ~3–5 minutes (TensorFlow is large).
python app.pyhttp://127.0.0.1:5000
💡 On first startup, models are trained automatically (~20–30 seconds). Subsequent starts load saved models instantly.
git init
git add .
git commit -m "AI Study Pal capstone"
git remote add origin https://github.com/YOUR_USERNAME/ai-study-pal.git
git push -u origin main- Go to https://render.com → Sign up free
- Click New → Web Service
- Connect your GitHub repo
- Render auto-detects
render.yaml— click Apply - Wait ~5 minutes for the first build
https://ai-study-pal.onrender.com
Share this URL with anyone worldwide — no installation needed.
48 hand-crafted questions across 6 subjects: Biology · Mathematics · History · Python · Physics · Chemistry
Each question has: subject, topic, difficulty (easy/medium/hard), 4 options, correct answer, and explanation.
Stored in ai_engine.py as QUESTIONS_DATASET — no external files needed.
The app prints ML metrics to the console on first startup:
📊 Quiz LR → Accuracy: 0.75 | F1 (weighted): 0.73
- Cleaned dataset (48 questions, Pandas DataFrame)
- Simple EDA visualisation (subject pie chart)
- ML-based quiz generator (TF-IDF + Logistic Regression + K-means)
- DL text summariser (Keras Dense NN)
- NLP study tips (NLTK keyword extraction)
- Motivational feedback (Keras Embedding NN)
- Flask web app with all 6 features
- Downloadable CSV study schedules
- Matplotlib charts in-browser
- Render.com deployment config
| Problem | Fix |
|---|---|
ModuleNotFoundError |
Run pip install -r requirements.txt |
| Slow first startup | Normal — models are training. Wait 30s |
| Port 5000 in use | Run python app.py — it auto-picks $PORT |
| TensorFlow install fails | Try pip install tensorflow-cpu==2.15.0 |
| NLTK download error | App auto-downloads to /tmp/nltk_data |
- AI-Generated Schedule - Creates personalized study plans based on exam date and available hours
- CSV Export - Download your schedule as a spreadsheet
- Pandas Integration - Intelligent date range distribution
- ML-Powered Questions - TF-IDF vectorization + Logistic Regression
- Difficulty Levels - Easy, Medium, Hard classifications
- Performance Metrics - Real-time accuracy tracking
- Deep Learning - Keras neural network with embedding layers
- Smart Extraction - Identifies most important sentences
- Quick Learning - Perfect for revision before exams
- NLP-Driven - NLTK-based keyword extraction
- Context-Aware - Tailored suggestions based on topics
- Personalized - Different tips for different subjects
- ML Clustering - K-means algorithm groups similar resources
- Smart Recommendations - Discover related learning materials
- Cross-Subject - Find connections between topics
- Neural Network - Keras embedding + classification
- Performance-Based - Adjusts tone based on quiz scores
- Encouraging - Boosts confidence for continued learning
User Input → Flask Route → AI Engine → ML/DL Model → JSON Response → UI Update
Example Flow (Quiz Generator):
- User selects a topic
- Flask receives POST request
ai_engine.pyloads the trained classifier- TF-IDF vectorizes the query
- Logistic Regression predicts difficulty
- App returns curated questions
- Frontend displays interactively
- Framework: Flask 3.0.0
- Data: Pandas 2.1.4, NumPy 1.26.4
- ML: scikit-learn 1.3.2
- DL: TensorFlow 2.16.1, Keras
- NLP: NLTK 3.8.1
- Plots: Matplotlib 3.8.2
- Server: Gunicorn 21.2.0
- Pure HTML5 + CSS3 + JavaScript
- No framework dependencies (Vue, React, Angular)
- Dark academic theme
- Fully responsive design
- Local: Flask dev server
- Production: Gunicorn WSGI
- Cloud: Render.com (free tier)
- Version Control: Git + GitHub
- Lines of Code: ~2,500+
- ML Models: 6 (various types)
- API Endpoints: 8
- Components: Modular architecture
- Test Coverage: Manual testing on all features
- Performance: <500ms for most requests
- User authentication (Firebase)
- Progress tracking & statistics
- Spaced repetition algorithm
- Mobile app (React Native)
- Video recommendations
- Group study features
- Real-time collaboration
- Advanced analytics dashboard
- Better ML models with more training data
- Transformer-based NLP (BERT)
- Distributed training on larger datasets
- Advanced difficulty adaptation
- Real-time progress visualization
Vishal Kothar
- GitHub: @ram0657-dev
- Email: ram0657@gmail.com
- Project: LaunchED Global Capstone (January 2026)
- Focus: Full-stack AI development with production deployment
✅ Python - Built complete backend application
✅ Machine Learning - TF-IDF, Logistic Regression, K-means
✅ Deep Learning - Neural networks for text summarization & feedback
✅ NLP - NLTK tokenization, keyword extraction
✅ Web Development - Flask, HTML/CSS/JavaScript
✅ Deployment - Production-ready Render.com deployment
✅ Git & GitHub - Version control & collaboration
✅ Data Processing - Pandas, NumPy, data cleaning
MIT License - Feel free to fork, modify, and use for your projects!
Found a bug? Have an idea?
- Fork the repository
- Create a feature branch
- Commit your changes
- Push and create a Pull Request
- Report issues in the GitHub Issues section
- 📧 Email: ram0657@gmail.com
- 🐛 Issues: GitHub Issues
- 💬 Discussions: GitHub Discussions
- Flask Documentation: https://flask.palletsprojects.com
- scikit-learn Guide: https://scikit-learn.org
- TensorFlow/Keras: https://www.tensorflow.org
- NLTK Handbook: https://www.nltk.org/book/
- LaunchED Global: Capstone curriculum
Built with passion for education and learning. Help students achieve their potential! 🚀
Last Updated: March 6, 2026