高级/资深Speech算法工程师（AIGC方向）

Excited to grow your career?

We value our talented employees, and whenever possible strive to help one of our associates grow professionally before recruiting new talent to our open positions. If you think the open position you see is right for you, we encourage you to apply!

Our people make all the difference in our success.

• Cutting-edge Technology: Focus on speech large language models and next-generation ASR technologies, driving innovation in speech understanding and generation.
• Core Impact: Direct contribution to company's flagship products in speech intelligence and audio processing platforms.
• Expert Collaboration: Work alongside world-class researchers and engineers in speech AI and large language models.
• Comprehensive Coverage: Research spans ASR, speech generation, speaker diarization, audio codecs, and speech-language model integration.

• Design and develop advanced speech algorithms based on large-scale models (Transformers, Conformers, Whisper-like architectures, and Speech LLMs).
• Lead or contribute to R&D and deployment in the following key areas:
• Automatic Speech Recognition (ASR): Develop robust end-to-end ASR systems for multilingual and multi-accent scenarios; optimize streaming ASR with ultra-low latency; implement context-aware and personalized speech recognition.
• Speech Generation: Advance zero-shot TTS with speaker adaptation; develop controllable speech synthesis with emotion and prosody modeling; create voice conversion and cross-lingual speech generation systems.
• Speaker Diarization: Build state-of-the-art speaker diarization systems for multi-speaker scenarios; develop joint ASR and diarization models; implement real-time speaker tracking and identification.
• Speech Large Language Models: Design and train speech-text multimodal LLMs; develop speech understanding models with reasoning capabilities; create unified models for multiple speech tasks.
• Neural Audio Codecs: Develop high-fidelity neural audio codecs for ultra-low bitrate transmission; optimize codecs for speech-specific applications; implement learnable compression for speech features.
• Stay current with academic and industry advances, driving adoption of breakthrough technologies into production systems.
• Oversee model training pipelines, performance optimization, and deployment strategies for large-scale speech models.

• Ph.D. in Computer Science, Electrical Engineering, Speech Processing, or related fields.
• Strong theoretical foundation in speech signal processing and deep expertise in deep learning, particularly in sequence modeling and generative models.
• Proficiency in Python and expertise in at least one deep learning framework (PyTorch, TensorFlow, JAX).

• Publications in top-tier venues such as INTERSPEECH, ICASSP, NeurIPS, ICML, ACL, IEEE TASLP, or Computer Speech & Language.
• Hands-on experience with large-scale model training, including speech foundation models, multimodal LLMs, or self-supervised speech models.
• Proven track record in deploying speech AI systems at scale, with experience in model compression and edge deployment.
• Experience with speech corpus creation, annotation, and quality assessment for large-scale training.