PeopleofSuperintelligence

Researcher

Sung-Feng Huang

Also known as 黃淞楓

Sung-Feng Huang is a speech and audio researcher whose work covers text-to-speech, audio representation learning and large audio language models. He studied electrical engineering at National Taiwan University from 2013 and completed a Ph.D. there in 2024 in the NTU Speech Processing and Machine Learning Lab, co-advised by Lin-shan Lee and Hung-yi Lee. His Meta-TTS paper of 2022 applied meta-learning to few-shot speaker adaptation, and in 2020 he coauthored a study of how ALBERT learns during pretraining. After interning at NVIDIA Research Taiwan from 2023 to 2024, he joined the group in March 2025. He is a Research Scientist there, working on generative AI for speech and audio.

Now at
NVIDIA
Works
39
Citations
255
h-index
8
Badges

Publication numbers from OpenAlex, as of 22 October 2026.

Timeline

  1. 2013

    Asian Physics Olympiad prizes

    At the 13th Asian Physics Olympiad in 2013, he wins a bronze prize and the Best Researcher Award.

  2. 2013–2017

    Electrical engineering at NTU

    In 2013, he enters National Taiwan University's electrical engineering department, and in 2017 he earns a B.S.

  3. 2017–2024

    Ph.D. in NTU's speech lab

    Ph.D. studies begin in 2017 at National Taiwan University, in the NTU Speech Processing and Machine Learning Lab. Lin-shan Lee and Hung-yi Lee co-advise him, and he completes the degree in 2024.

  4. 2019

    Audio Word2vec in TASLP

    In 2019, IEEE/ACM TASLP publishes Audio Word2vec, a sequence-to-sequence autoencoding method for unsupervised learning of audio segmentation and representation, with him as second author.

  5. 2020

    A study of ALBERT pretraining

    With Cheng-Han Chiang and Hung-yi Lee, he coauthors Pretrained Language Model Embryology: The Birth of ALBERT, accepted to EMNLP 2020. The paper reports that ALBERT learns to reconstruct and predict tokens of different parts of speech at different speeds during pretraining.

  6. 2022

    Meta-TTS for few-shot voices

    As first author, he proposes Meta-TTS, published in IEEE/ACM Transactions on Audio, Speech, and Language Processing in 2022. It trains a multi-speaker text-to-speech model with Model Agnostic Meta-Learning so the model adapts quickly to a new speaker from a few enrollment samples.

  7. 2023–2024

    NVIDIA internship and dissertation

    He works as an AI researcher intern at NVIDIA Research Taiwan from October 2023 to August 2024. His 2024 National Taiwan University dissertation, Personalized Text-to-Speech Synthesis: Lightweight and Fast Few-Shot Adaptation for Unseen Speakers, Accents and Languages, covers Meta-TTS, adaptive structured pruning and few-shot cross-lingual TTS.

  8. Since March 2025

    Research Scientist at NVIDIA

    He joins NVIDIA Research Taiwan as a Research Scientist in March 2025, working on generative AI for speech and audio, including speech recognition, synthesis and separation. In July 2025, he is among 28 authors who introduce DeSTA2.5-Audio, a general-purpose large audio language model.

Career

  1. Since 2023NVIDIAAI Researcher Intern, Research Scientist

Education

  1. 2024National Taiwan UniversityPh.D.
  2. 2017National Taiwan UniversityB.S.

From the biography and OpenAlex affiliations (the years of papers that list the organization), so some years are approximate.

  1. 16% of AI career moves in the 2020s went from a university to a companyIncludes Sung-Feng Huang. We estimate that 16% of AI career moves in the 2020s went from a university to a company; 29% of moves starting at a university went to a company, and 76% ended the university role.
  2. Google has the highest combined h-index among AI organizationsIncludes NVIDIA. We estimate that Google holds the highest combined h-index at 8,016, while universities account for 73% of the deduplicated total across AI organizations.
  3. 20% of people at companies who shape AI also hold a university role, led by Google and Stanford UniversityIncludes NVIDIA. We estimate that 539 of the 6,766 people with a known current organization (8%) hold both a university role and a company role, including 20% of people at companies and 13% of people at universities.
  4. Google alumni have founded or led 133 other organizations, more than any other AI labIncludes NVIDIA. We estimate that alumni of 17 major AI labs went on to found or lead 350 other organizations, including 226 they founded, and we trace 472 career moves directly between 23 labs since 2010.

Colleagues

People named in this biography, and people with dated positions at the same organization in the same years.

Sources and credits

Photo: We have no published photo of Sung-Feng Huang yet. Send us a free-licensed photo.

  1. Sung-Feng Huang - Sung-Feng Huang sungfeng-huang.github.io. Checked 20 October 2026.
  2. Sung-Feng Huang | NVIDIA Research Taiwan research.nvidia.com. Checked 20 October 2026.
  3. NTU Theses and Dissertations Repository: 個人化文字轉語音：針對新語者、口音及語言的快速輕量化少樣本適應 tdr.lib.ntu.edu.tw. Checked 20 October 2026.
  4. [2010.02480] Pretrained Language Model Embryology: The Birth of ALBERT arxiv.org. Checked 20 October 2026.
  5. [2111.04040] Meta-TTS: Meta-Learning for Few-Shot Speaker Adaptive Text-to-Speech arxiv.org. Checked 20 October 2026.
  6. [2507.02768] DeSTA2.5-Audio: Toward General-Purpose Large Audio Language Model with Self-Generated Cross-Modal Alignment arxiv.org. Checked 20 October 2026.