Developed cross-provider speech capabilities for the ProjectTech4DevAI/ai-platform repository by integrating Speech-to-Text and Text-to-Speech features using Sarvam and ElevenLabs. Focused on a provider-agnostic backend architecture, the work enabled in-app speech recognition and synthesis, laying the foundation for expanded language support and improved accessibility. Leveraged Python for backend development, API integration, and data modeling, with unit testing ensuring reliability. The integration reduced future vendor lock-in and streamlined the addition of new providers. Collaboration was demonstrated through co-authored, well-structured commits, and the delivered feature positioned the platform for flexible, multilingual voice-enabled user experiences without introducing new bugs.