Jaeseung Lee
Research Engineer
Digital Pathology · Machine Learning
New York, NY
I develop machine learning and data analysis pipelines for digital pathology and translational biomedical research. My work includes machine learning, whole-slide image (WSI) analysis and spatial transcriptomics.
I am interested in building practical tools that connect computational methods with pathology and improve clinical research workflows.
Email: jaeseunglee124 [at] gmail [dot] com
About
My background is in electrical engineering and machine learning. I have worked on computational pathology projects involving deep learning-based whole-slide image (WSI) classification, tissue segmentation, and spatial analysis of transcriptomics data.
Previously, I contributed to research at Columbia University Irving Medical Center and Korea University of Technology and Education, where I developed machine learning models and scalable analysis pipelines.
Professional Experience
- Developed a GPU-accelerated analysis pipeline for spatial transcriptomics to quantify cell–cell interactions, achieving up to 1200× speedup; co-authored a publication and released the code on GitHub.
- Built a deep learning framework for WSI analysis to classify Huntington’s disease vs controls in a 66-case cohort, achieving 95% accuracy and generating attention heatmaps for interpretable feature localization.
- Developed a tissue segmentation pipeline using Reinhard normalization to reduce stain variation, achieving 97% accuracy in multi-class tissue classification.
- Maintained RNA-seq data processing workflows for RNA-seq and spatial transcriptomics, including preprocessing, alignment, and quality control.
- Developed a machine learning system to detect physiological responses to music using wearable sensor data and deployed trained models in a mobile application and achieved 81% prediction accuracy; Co-authored a peer-reviewed publication based on this work.
Skill
- Programming: Python, C/C++, Bash, Linux, Git/GitHub
- Machine Learning: PyTorch, TensorFlow, Scikit-learn
- Databases & Web App: MongoDB, Neo4j, HTML/CSS, JavaScript
- Digital Pathology: OpenSlide, QuPath, ImageJ/Fiji
Publication
-
Ma, M., Paryani, F., Jakubiak, K., Xia, S., Antoku, S., Kannan, A., Lee, J., ... , Sims, P. & Al-Dalahmah, O.
The spatial landscape of glial pathology and T cell response in Parkinson’s disease substantia nigra.
Nature Communications. 2025; 16:7146. [Paper] [GitHub] -
Lee, E., Min, C., Lee, J., Yu, J. and Kang, S.
Groovemeter: Enabling music engagement-aware apps by detecting reactions to daily music listening via Earable sensing.
Proceedings of the 31st ACM International Conference on Multimedia. 2023. [Paper]