This project focuses on the downstream NLP task of Machine Translation in the legal domain. It aims to develop and evaluate models capable of accurately translating legal documents between languages while preserving meaning, context, and formal structure. By leveraging domain-specific data and fine-tuning transformer-based architectures, the project addresses the unique challenges of legal language, including complex sentence structures, jargon, and jurisdictional nuances.