Research Scientist Intern, Speech/Audio Generation

Responsibilities:

• Conduct cutting-edge research and development in speech/audio foundation models.
• Research, design, develop, and evaluate generative models.
• Conduct research to integrate Large Language Modeling into speech and audio generation (Neural Tokenizers, Large Language Modeling, etc.).
• Identify key research areas and contribute to innovative speech/audio models.
• Design and curate large-scale, high-quality datasets with multi-team efforts.

Qualifications:

• Major in computer science, mathematics, engineering, a related field, or equivalent professional experience.
• 2+ years of internship or full-time experience in one or more areas of machine learning and deep learning, including but not limited to:Speech/Vocal generation (including text-to-speech, singing-voice-synthesis, prosody transfer, etc.)
  Audio generation (including text-to-music, etc.)
  Large scale speech/audio self-supervised representation learning and foundation models
  Large Language Model pre-training and fine-tuning
• Deep knowledge of deep learning and generative models (Diffusion, Flow Matching, AR Transformer, Mamba, VAE, GAN, etc.).
• Self-driven, innovative, collaborative, with strong communication and presentation skills.

Preferred Qualifications:

• Ability to work collaboratively in fast-paced, startup environment.
• Strong publications in top-tier ML/Speech conferences/journals (e.g., NeurIPS, ICML, ICLR, ICASSP, INTERSPEECH).
• Deep understanding of Large Language Models.
• Familiarity with different variants of diffusion models, such as Flow matching, DDIM, etc.
• Familiarity with large-scale distributed training.
• Knowledge of engineering principles and best practices.
• High proficiency in algorithms and programming; strong coding skills in Python.