Paper Title
Nyay Sahayak: AI-Based Virtual Judge System

Abstract
This paper presents an AI-based Legal Judgment Prediction system that assists in analyzing legal cases using Retrieval-Augmented Generation (RAG) and machine learning techniques. The system processes unstructured legal text, summarizes case facts, retrieves relevant precedents, and predicts outcomes as Plaintiff or Defendant win. To address class imbalance, SMOTE is applied, and explainable AI techniques ensure transparency. Experimental results demonstrate a balanced F1-score of 0.855, indicating effective and reliable performance. The proposed system improves legal research efficiency and supports data-driven judicial decision-making. Keywords - Legal Judgment Prediction, Retrieval-Augmented Generation, Large Language Models, Natural Language Processing, Legal Tech, Two-class Classification, Interpretable AI, Case Law Matching.