Passionate about Data Science | Research Intern @ Samsung R&D | Continuously Exploring and Learning the World of Data
Email: Pranaykarvi@gmail.com | Phone: +91 8655287793 | Resume: View Resume
• HDR Algorithm Development: Designing and evaluating exposure-bracketing algorithms to maximize dynamic range while minimizing noise under real-time constraints
• Research & Implementation: Conducting literature review and implementing hybrid pipelines combining image processing, multi-scale histogram analysis, and neural networks
• Model Optimization: Developing and optimizing models using Python, OpenCV, NumPy, PyTorch, and TensorFlow with focus on latency–quality trade-offs
• Testing & Validation: Performing robustness testing across diverse lighting scenarios and datasets
• Documentation: Preparing technical documentation and contributing to research paper drafting
• Alzheimer's Detection: Developed deep learning pipelines for disease detection using MRI and PET multimodal data
• CNN Architectures: Implemented and benchmarked VGG16, Xception, DenseNet201, and EfficientNet-B3
• Explainable AI: Applied SHAP and LIME techniques to improve clinical interpretability
• Data Pipeline: Worked on preprocessing, augmentation, and multi-modal feature fusion
• Publication: Research manuscript submitted and under review (IEEE) | View Paper
Project: RIMUS (Real-time Intelligent Multi-Utility System)
• Autonomous Rover: Designed industrial safety rover with sensor-first, fail-safe system integration
• Hardware Integration: Integrated ESP32 sensing, edge decision logic, vision/thermal perception
• Real-time Dashboard: Built real-time monitoring dashboards
• Team: Harsh Singhal, Aqeeb Akeel, Lokesh Maan, Niket Suthar | Mentor: Dr. Ravi Prakash Dwivedi
• Healthcare Ecosystem: Built unified digital healthcare platform for rural accessibility
• Team: Harsh Singhal, Lokesh Maan, Niket Suthar, Madhumitha, Vani Makhija
• Mentors: Dr. Sukriti, Dr. Suganthi
| Hackathon | Achievement | Repository |
|---|---|---|
| Google GenAI Hackathon | AI-powered deepfake and fake news detection | Repository |
| Bajaj HackRx | Hackathon project submission | Repository |
| Welldoc Hackathon | Built explainable diabetes risk prediction (XGBoost + SHAP) | Repository |
| Project | Description | Tech Stack | Links |
|---|---|---|---|
| Derma-AI | Skin lesion classification using ViT + EfficientNet fusion | Next.js TypeScript ViT EfficientNet |
  |
| Pneunomia-AI | Ensemble-based pneumonia detection from chest X-rays | Deep Learning Ensemble CNN Streamlit |
 |
| Plant Disease Classifier | CNN-based model for plant disease detection | CNN Deep Learning Streamlit |
|
| Project | Description | Tech Stack | Links |
|---|---|---|---|
| ResearchMate | AI-powered research assistant Chrome Extension | Chrome Extension LangChain FastAPI |
 |
| Prepy-AI | AI career assistant for resumes, cover letters & interviews | Next.js Clerk Neon Gemini API |
|
| Smart AI Assistant | Voice-controlled AI assistant with modular agents | Python Voice Recognition Modular Agents |
|
| Project | Description | Tech Stack | Links |
|---|---|---|---|
| Apple Stock Predictor | LSTM & GRU-based time-series forecasting model | LSTM GRU Time Series TensorFlow |
|
| Topography Classifier | Satellite image classification into land types | CNN Computer Vision Satellite Imagery |
|
| Accident Emergency Responder | Real-time emergency detection using geolocation & ML | ML Geolocation Real-time Flask |
|
| Organization | Contribution | Pull Requests |
|---|---|---|
 |
Improved CI test assertions and documentation | PR #5772 • PR #5718 |
 |
Migrated test framework to Pytest; improved scalability | PR #2451 • PR #2561 • PR #2580 |
 |
Refactored condition update logic | PR #8002 |
 |
Improved logging practices | PR #6400 |
 |
Distributed training enhancements | PR #1531 |
 |
Bug fixes and CLI usability improvements | PR #7983 |
• Workshops & Events: Led technical workshops, coding sessions, and 4 large-scale hackathons (250–300+ participants each)
• Mentorship: Mentored 55+ members in programming, ML, and research practices
• Research Support: Supported research paper writing and publication efforts
• Growth: Increased club membership by ~30% through community initiatives
• Managed budgeting, sponsorships, and outreach for technical events
• Contributed to AI projects and technical workshops
