A collection of independent deep-dives into AI, Healthcare, and Developer Tools.
Hand-picked projects representing best solo work.
|
AI-Powered Adaptive Interview Preparation Platform
Adaptive interview preparation platform with AI-driven personalized learning paths. System analyzes user performance across 16 core algorithmic patterns and adjusts difficulty dynamically.
| Feature |
Implementation |
| Adaptive Learning Engine |
Performance-based difficulty scaling with personalized question curation |
| Real-time AI Coaching |
GPT-4 powered hints and solution walkthroughs |
| Progress Analytics |
Detailed performance breakdown by algorithmic pattern |
| DSA Mastery Paths |
Structured curriculum covering 16 fundamental patterns |
Tech Stack:
- Frontend: Next.js 15, React, TypeScript, Tailwind CSS, Framer Motion
- Backend: FastAPI, Python, AsyncIO
- AI: OpenAI GPT-4 API, custom prompt engineering
- Database: PostgreSQL for user progress tracking
Links: GitHub
|
|
AI-Powered Algorithmic Interview Mastery Platform
Comprehensive interview preparation platform mastering 16 core algorithmic patterns with real-time AI coaching. Provides structured learning paths from fundamental arrays to advanced dynamic programming.
| Module |
Description |
| Pattern Recognition |
Curated problem sets organized by 16 core algorithmic patterns |
| Live AI Assistance |
Context-aware hints and solution explanations |
| Code Execution |
Integrated sandbox environment with test case validation |
| Progress Tracking |
Visual analytics on mastery level per pattern |
Tech Stack:
- Frontend: Next.js 15, React, TypeScript, Tailwind CSS
- AI: OpenAI GPT-4, custom coaching prompts
- Execution: Docker-based code sandbox
Links: GitHub
|
|
AI-Powered Medical Imaging for Bone Fracture Detection
Medical imaging platform for automated bone fracture detection in X-rays using computer vision. Deploys YOLOv8 object detection pipeline optimized for radiological analysis with sub-second inference.
| Capability |
Technology |
| Fracture Detection |
YOLOv8 fine-tuned on medical imaging datasets |
| Localization |
Bounding box prediction with confidence scores |
| Report Generation |
Automated preliminary findings documentation |
| DICOM Support |
Medical image standard compatibility |
Tech Stack:
- Model: YOLOv8, PyTorch, custom medical dataset training
- Backend: FastAPI, Python, OpenCV
- Frontend: React, TypeScript, medical imaging viewer components
- Deployment: Docker containerization
Links: GitHub
|
|
AI-Powered Personalized Nutrition Recommendation Engine
Personalized diet recommendation platform using Gemini AI for tailored meal plans. Analyzes user goals, dietary restrictions, and preferences to generate nutritionally balanced plans with macro tracking.
| Feature |
Implementation |
| AI Meal Planning |
Gemini-powered daily/weekly plan generation |
| Macro Tracking |
Protein, carb, fat breakdown with calorie counting |
| Dietary Restrictions |
Vegan, keto, gluten-free, allergy-aware planning |
| Recipe Database |
Curated healthy recipes with nutritional info |
Tech Stack:
- Frontend: Next.js 15, React, TypeScript, Tailwind CSS
- AI: Google Gemini API, nutritional prompt engineering
- State: Local storage with export capabilities
Links: GitHub
|
|
Cognitive Skills Assessment Platform
Interactive cognitive skills testing platform with gamified assessments for memory, reaction time, and problem-solving. Provides detailed analytics on cognitive performance metrics.
| Assessment Type |
Description |
| Memory Tests |
Pattern recognition and recall challenges |
| Reaction Time |
Visual and auditory stimulus response measurement |
| Problem Solving |
Logic puzzles and pattern matching |
| Progress Tracking |
Historical performance comparison |
Tech Stack:
- Frontend: Next.js 15, React, TypeScript, Tailwind CSS
- Animation: Framer Motion for interactive elements
- Deployment: Vercel Edge Network
Links: Live Demo | GitHub
|
|
Hackathon Discovery and Management Platform
Platform aggregating hackathon events from Devfolio and Devpost with intelligent filtering and team formation features. Helps developers discover relevant competitions and manage submissions.
| Feature |
Implementation |
| Event Aggregation |
Web scraping and API integration from major platforms |
| Smart Filtering |
Tech stack, location, prize pool, deadline filters |
| Team Builder |
Profile-based teammate matching |
| Bookmarking |
Save and track interesting hackathons |
Tech Stack:
- Frontend: Next.js 15, React, TypeScript, Tailwind CSS
- Backend: FastAPI, Python, AsyncIO
- Database: PostgreSQL with full-text search
- Scraping: BeautifulSoup, Selenium for data aggregation
Links: Live Demo | GitHub
|
|
Machine Learning Pipeline for Financial Fraud Detection
End-to-end fraud detection pipeline for IEEE-CIS Kaggle competition. Implements advanced feature engineering and ensemble methods to identify fraudulent transactions with high precision.
| Component |
Approach |
| Feature Engineering |
Transaction aggregation, time-based features, device fingerprinting |
| Model Ensemble |
XGBoost, LightGBM, CatBoost stacking |
| Imbalance Handling |
SMOTE, class weighting, focal loss |
| Evaluation |
ROC-AUC, Precision-Recall optimization |
Tech Stack:
- ML: XGBoost, LightGBM, CatBoost, Scikit-learn
- Processing: Pandas, NumPy, feature-engine
- Validation: Stratified K-fold, time-series split
Links: GitHub
|
|
Personal Portfolio Website
Personal portfolio showcasing full-stack development and AI specialization. Features smooth animations, project showcases, and responsive design with modern UI patterns.
| Section |
Features |
| Hero |
Animated introduction with typewriter effects |
| Projects |
Interactive project cards with live demos |
| Skills |
Visual skill proficiency indicators |
| Contact |
Integrated contact form with validation |
Tech Stack:
- Frontend: Next.js 15, React, TypeScript, Tailwind CSS
- Animation: Framer Motion, scroll-triggered animations
- Icons: Lucide React
- Deployment: Vercel
Links: Live Site | GitHub
|
