B.Tech AI & Data Science · Amrita Vishwa Vidyapeetham, Coimbatore
I build AI systems that actually run — not just notebooks. My work spans deep learning, computer vision, NLP, and agentic AI workflows. Two of my projects have associated research publications.
Audio Deepfake Detection — CNN-LSTM model on MFCC features. Research paper accepted at IEEE CONIT 2026.
Crowd Monitoring AI — YOLOv8 + DeepSORT real-time crowd safety system. Springer journal paper under review.
Agentic AI — LangGraph-based multi-turn agents with RAG, intent detection, and LLM integration.
Computer Vision — Real-time drowsiness detection, PCB defect detection, crowd density analysis.
| Project | What it does | Stack |
|---|---|---|
| VoxGuard | Audio deepfake voice detection (CNN-LSTM) | Python, TensorFlow, Librosa |
| CrowdSentry | Real-time crowd monitoring + alert dashboard | Python, YOLOv8, Streamlit |
| AutoStream Lead Agent | Conversational AI agent for lead capture | LangGraph, Gemini, Groq |
| DrowsyGuard | Driver drowsiness detection via webcam | Python, OpenCV, CNN |
| ResumeRank | Resume screening with TF-IDF + BERT | Python, Scikit-learn, BERT |
| GridSense | LSTM energy forecasting for microgrids | Python, TensorFlow, LSTM |
- "Robust Deepfake Audio Detection Using Temporal Deep Learning" — IEEE CONIT 2026 (Accepted)
- "CrowdSentry: Intelligent Cloud-Integrated Crowd Monitoring" — Springer Journal (Under Review)
Python C++ TensorFlow OpenCV LangGraph LangChain Scikit-learn React FastAPI Streamlit Git