Yueh-Po Peng
|
|
|
Experience
AI Engineer Oct 2024 – Present
Gate.io Taipei, Taiwan (Remote)
Developed a fund flows anomaly detection system with LLMs and tree-based models, enhancing financial security.
Built a Text-to-SQL system to streamline internal data queries, improving query efficiency by 20%.
Research Assistant Mar 2022 – Oct 2024
Institute of Information Science, Academia Sinica | MCTLAB | Supervisor: Dr. Li Su
Taipei, Taiwan
Research Topics: Self‑Supervised Learning, Medical Imaging
Proposed a Transformer-based self-supervised learning method for decoding brain signals (fMRI), achieving an 77% reduction in memory footprint.
Conducted distributed training experiments on high-resolution 4D medical images (fMRI) using TWCC HPC.
Proposed a whole-brain feature selection method for decoding musical pitch from fMRI [2] .
AI Engineer Intern Mar 2023 – Jul 2024
Tomofun - World's leading pet technology company Taipei, Taiwan
Research Topics: Computer Vision, Large Language Models, Multimodal Learning
Developed an automatic short music video generation system for daily pet clips.
Fine-tuned visual language models (e.g., BLIP), achieving a 20.6% improvement in visual question answering.
Enhanced LLaVA image inference speed by 250% with only a 3% accuracy reduction.
Developed APIs for visual language models using llama.cpp/ollama for image-caption pair datasets.
Education
National Taiwan University Feb 2023 – Jun 2024
Taipei, Taiwan
M.S. in Data Science
Thesis topic: Whole‑Brain Feature Selection Methods for Decoding from fMRI Data
National Taiwan University Sep 2019 – Jan 2022
Taipei, Taiwan
B.S. in Computer Science and Information Engineering (CSIE)
Research & Projects
Guitar Effect Removal Collaboration with Positive Grid ML Team
Proposed a two-stage method to remove distortion effects from guitar recordings using Positive Grid VST plugins.
Achieved 20% higher audio quality than the best baseline, rated by 26 professional guitarists.
Published in DAFx 2024 [1] .
Whole Brain fMRI Feature Selection
Proposed a two-stage method to extract fMRI features and predict musical pitch.
Demonstrated 2-fold improvement over ROI-based feature selection in fMRI-music analysis.
Published in ICASSP 2023 [2] .
Publications
[1] Lee, Y. S.*, Peng, Y. P.* , Wu, J. T., Cheng, M., Su, L., & Yang, Y. H.
"Distortion Recovery: A Two-Stage Method for Guitar Effect Removal," Proc. Int. Conf. Digital Audio Effects 2024 (DAFx’24). (* equally contributed)
Paper | Demo
[2] Cheung, V. K.*, Peng, Y. P.* , Lin, J. H., & Su, L.
"Decoding Musical Pitch from Human Brain Activity with Automatic Voxel-Wise Whole-Brain FMRI Feature Selection," Proc. IEEE Int. Conf. on Acoustics, Speech, and Signal Processing 2023 (ICASSP’23). (* equally contributed)
Paper
Skills
Languages/Frameworks : Python, PyTorch, TensorFlow, Pandas, Scikit-learn, Slurm, Go, HTML, JavaScript, C++, C, Linux.
Skillset : Self-Supervised Learning, Medical Imaging, Computer Vision, Music Information Research, Distributed Training.