- Conducting research and implementing state-of-the-art hybrid (Kaldi) and deep learning (PyTorch, k2) models to develop a robust ASR system for an AI Voice Assistant. Successfully achieving high accuracy in real-world environments.- Developing and implementing a multilingual (Vietnamese + English) data processing pipeline to support the ASR model's training and evaluation.- Implement a streaming ASR model with low latency in the production environment.- Continuously analyzing and optimizing the ASR model's performance in production.