The role focuses on researching and developing state-of-the-art TTS models based on LLMs and Flow Matching for next-generation mobility solutions
Job Summary
The role focuses on researching and developing state-of-the-art TTS models based on LLMs and Flow Matching for next-generation mobility solutions.
Candidates must possess strong experience in optimizing models for both server and on-device environments to ensure low latency and high quality.
42dot is a global software center of Hyundai Motor Group dedicated to leading software-defined vehicle development and autonomous urban traffic operations.
Matching Summary
The role focuses on researching and developing state-of-the-art TTS models based on LLMs and Flow Matching for next-generation mobility solutions.
Skills & Requirements
Must-have
3+ years TTS development experience
Multilingual and multi-speaker model development
Model optimization and quantization expertise
TensorRT and ONNX based optimization
Linux environment development skills
Python and C++ programming proficiency
Nice-to-have
Neural Audio Codec research experience
Streaming TTS system development background
Flow Matching and Diffusion model knowledge
Top-tier conference paper publications
Instruction-based Text LLM experience
Key Requirements
Minimum 3 years of TTS related career experience
Master's or PhD degree in relevant fields preferred
Experience with PyTorch and open-source deep learning frameworks