Computer Science Undergraduate | AI & ML Enthusiast | Computer Vision & Data Science Skilled
I'm a AI & ML enthusiast and data science student passionate about Deep Learning, Computer Vision, and LLM. My mission: to build intelligent and end-to-end real world solutions like integrated parking systems to agentic RAG platforms and YOLO & Machine Learning Models
Current Focus: AI/ML Engineering • Data Science & Analytics • MLOps • RAG • LLM
Supervised/unsupervised learning, neural networks, fine-tuning
Real-time detection, image classification, biometric analysis
Agentic frameworks, vector databases, LLM integration
EDA, feature engineering, time-series forecasting
Model deployment, system architecture, API development
Real-time smart parking platform integrating YOLOv8 for sub-second slot detection and XGBoost for time-series availability forecasting. Features cross-platform Flutter mobile app with last-mile navigation and robust Flask backend with Streamlit dashboard.
Tech Stack: Python, YOLOv8, PyTorch, OpenCV, Dart, SQL, XGBoost, REST API, Firebase, GeoFire, Google Maps SDK, Flutter, Streamlit
An intelligent PDF assistant driven by Multimodal RAG and Groq's ultra-fast LLM inference that enables users to chat with their PDF documents and automatically generate comprehensive exam questions from study materials.
Tech Stack: RAG, Python, LangChain, FAISS, Groq API (Llama-3), SentenceTransformers, Streamlit, PyPDF, ReportLab
A cutting-edge Riemannian framework for safe machine unlearning. Uses the Fisher-Rao metric to guide parameters along the "safe" data manifold, strictly preserving utility on remaining data while surgically erasing target classes via adversarial optimization. Superior to Euclidean L2 baselines.
Tech Stack: Python, PyTorch, Differential Geometry, Optimization
ScribeScan is a sophisticated web application that leverages deep learning to automatically identify student IDs from handwritten document submissions. It bridges the gap between analog handwritten assessments and digital record-keeping, enabling educators to instantly attribute authorship to handwritten papers without manual verification.
Tech Stack: Python, TensorFlow, Keras, OpenCV, Flask
- 🏆 Hult Prize National Finalist - AquaSense Project (2025)
- 🤖 Building agentic RAG systems & real-time AI applications
- 🧠 Advanced Diffusion Models & Generative AI
- 🌐 Domain Generalization in Computer Vision
- 📡 Large-Scale MLOps & Model Deployment
- 🔗 Multi-Agent AI Systems & Orchestration
