I'm an AI Researcher and Engineer passionate about building intelligent systems at the intersection of Computer Vision, Generative AI and Reinforcement Learning.
Currently, I am:
- 🔬 Research Intern at Reagvis Labs
- 🎓 Summer Research Intern at IIT Mandi
My work focuses on developing AI systems that solve real-world problems in healthcare, security, and autonomous decision-making.
I am actively working on projects involving:
- Medical AI
- Large Language Models (LLMs)
- Reinforcement Learning
- Deepfake Detection & Localization
- Agentic AI Systems
- Autonomous Security Agents
An end-to-end Medical Assistant Chatbot built using LangChain and Pinecone Vector Database.
MedGPT leverages Retrieval-Augmented Generation (RAG) to answer medical queries using trusted medical knowledge sources such as the Gale Encyclopedia of Medicine. The platform provides an intuitive conversational experience and integrates both backend intelligence and user-friendly interfaces.
MedTSRL: A Tutor-Student Reinforcement Learning Framework for Density-Aware Medical Deepfake Localization
A research-focused framework designed to improve localization of manipulated regions in medical images using reinforcement learning and density-aware supervision. MedTSRL aims to bridge the gap between accurate classification and precise manipulation localization in medical imaging systems.
WRAITH is a fully local, autonomous AI-powered penetration testing agent designed to perform intelligent security assessments.
The system leverages local Large Language Models to autonomously reason, plan, and execute penetration testing workflows against web targets while maintaining a completely local execution environment.
- Medical Artificial Intelligence
- Deepfake Detection
- Computer Vision
- Reinforcement Learning
- Agentic AI Systems
- Large Language Models
- AI for Cybersecurity
- Explainable AI
- GitHub: https://github.com/VectorCipher
- LinkedIn: https://linkedin.com/in/naitikpal
Building intelligent systems that combine research, autonomy, and real-world impact.