Multi-Language Speech Translation Prototype

I need a speech-to-speech prototype developed for real-time Mandarin to Hindi conversation. The system should ensure instant translation, accurate pronunciation, and incorporate noise cancellation.

Key Requirements:
- Offline functionality, portable on devices like Raspberry Pi
- Ability to add custom languages like Tibetan and Ladakhi
- Fine-tuning of the model for accuracy and efficiency

Ideal Skills and Experience:
- Expertise in speech recognition and synthesis
- Experience with real-time translation technologies
- Proficiency in machine learning and model optimization
- Familiarity with Raspberry Pi and offline deployment

Looking for a freelancer with a strong background in linguistics and AI.