Developed and integrated Kaapi Guardrails for input and output validation in the ProjectTech4DevAI/ai-platform repository, focusing on enhancing safety and compliance for large language model (LLM) operations. The work centered on filtering sensitive information during LLM calls, reducing data leakage risks and supporting enterprise governance requirements. Leveraging Python for backend development, the implementation utilized API integration and robust data validation techniques, with unit testing ensuring reliability and deployment readiness. The feature was delivered as a complete, commit-driven solution, with no major bugs reported, demonstrating a methodical approach to LLM safety engineering and compliance within a production-ready environment.