About Me

Hello! My name is Xiaoqin Feng. I am a Master's student at the University of Southern California (USC) with research interests in Generative AI, LLM-powered agents, Human-AI Collaboration, and AI for Science. My work emphasizes agents, speech, data, evaluation, and real-world applications. I bring hands-on industry experience in these areas and am dedicated to advancing artificial intelligence through rigorous research and practical innovation.

Here is my CV. I am actively seeking PhD programs starting in 2027 and 2026 Summer Internships (SDE/ML/AI Engineering, etc.)—feel free to email me if you're interested.

AI Products

I am building a next-generation AI audio visualization experience that blends speech intelligence with interactive visuals. You can try the latest demo and share feedback to shape where it goes next.

  • Live demo: qinzhi.org
  • Looking for collaborators, beta users, or partners—please reach out if this resonates with you.

Education

Employment and Activity

  • [Aug. 2025 - Present] Mobvoi AI Lab, Part-time Consultant. Support product decision-making by focusing on model optimization, multi-agent systems, and real-time information tracking.
  • [May. 2023 - Jul. 2025] Mobvoi AI Lab, Tech Lead. Research on LLM agent, LLM application, Audio LLM, prompt engineering, data engineering and model evaluation. (Agent & Application & Data & Evaluation)
  • [Jul. 2019 - May. 2023] Mobvoi AI Lab, Senior Speech Engineer. Research on speech synthesis on multi- and cross- lingual domain, including semantic NLP, emotional NLP and generative NLP.
  • [Aug. 2018 - Dec. 2018] TAL AI Lab, Algorithm Research Intern. Research on deep knowledge tracing and graph representation learning.
  • [Jun. 2018 - Aug. 2018] DeeCamp AI Lab, Research Team Member. Research on 'Movie Recommendation based on Knowledge Graph'. github
  • [Sep. 2017 – Dec. 2017] Ali Tianchi Competition, Competitor. Topic was predicting user's current store location based on user consumption data (one million+). Research on linear model, ensemble model, and big data mining.
  • waiting to be added .....

News

  • [Dec. 2025] Launched an interactive AI audio visualization web—try it here: qinzhi.org.
  • [Oct. 2024] Our paper about Audio LLM - SparkTTS is available on arXiv 2025.
  • [Oct. 2023] Our paper about information extraction and Multilingual has been accepted by APIN (Applied Intelligence) 2024.
  • [Oct. 2022] Our paper about multi-granularity stress prediction of expressive TTS has been accepted by APSIPA 2023.
  • [Jul. 2022] Our paper about prompt learning (rhythm) has been accepted by PRML 2022.
  • [Jun. 2019] Our paper about knowledge tracing has been accepted by AIED 2019.
  • [Jun. 2019] Our paper about image detection has been accepted by Journal of Physics 2019.
  • [Oct. 2018] Our paper about quick edge detection (image) has been accepted by CCIOT 2018.
  • [Nov. 2018] Our paper about cloud computing and intelligence system has been accepted by CCIS 2018.

Publications [Google Scholar]

  1. SparkTTS SparkTTS
    Spark-TTS: An Efficient LLM-Based Text-to-Speech Model with Single-Stream Decoupled Speech Tokens
    Xinsheng Wang, Xiaoqin Feng, et al.
    arXiv preprint 2025, 22 pages
    PDF
  2. APIN 2024 APIN
    2M-NER: Contrastive Learning for Multilingual and Multimodal NER with Language and Modal Fusion
    Dongsheng Wang, Xiaoqin Feng, et al.
    In proceedings of Applied Intelligence 2024, 20 pages
    PDF
  3. APSIPA 2023 APSIPA
    Multi-granularity Semantic and Acoustic Stress Prediction for Expressive TTS
    Wenjiang Chi, Xiaoqin Feng(*equal contribution), et al.
    In proceedings of APSIPA 2023, 5 pages
    PDF Demo
  4. PRML 2022 PRML
    Prosody Prediction with Discriminative Representation Method
    Jipeng Zhang, Hankiz Yilahun, Xiaoqin Feng, Yunlin Chen, Xipeng Yang, Askar Hamdulla
    In proceedings of PRML 2022, 5 pages
    PDF
  5. AIED 2019 AIED
    Deep knowledge tracing with side information
    Zhiwei Wang, Xiaoqin Feng, Jiliang Tang, Gale Yan Huang, Zitao Liu
    In proceedings of AIED 2019, Springer, Cham, 5 pages
    PDF
  6. Journal of Physics 2019 Journal of Physics
    Population statistics algorithm based on MobileNet
    Xiaoqin Feng, Rong Xie, Junyang Sheng, Shuo Zhang
    Journal of Physics: Conference Series. JPCS 2019, 6 pages
    PDF
  7. CCIOT 2018 CCIOT
    A method of quick edge detection based on Zynq
    Rong Xie, Xiaoqin Feng
    In proceedings of CCIOT 2018, 5 pages.
    PDF
  8. CCIS 2018 CCIS
    Research on the Internet of Things Platform for Smart and Environmental Protection
    Junyang Sheng, Xiaoqin Feng
    In proceedings of CCIS 2018, 5 pages
    PDF

Teaching

  • Lecture - Embedded System Design Practice
    As teaching assistant at BJUT, for M.Sc. students, approx. 80 students each year. Winter 2018
  • Company - Speech & NLP
    As a mentor at Mobvoi, for interns (students), avg. 3 students each year. Annual 2021 - 2025

Life Cakes

  • I love music, cooking, reading, traveling...
  • Here is my life story, expecting .....