Curriculum Vitae

Education

• Ph.D in Electrical Engineering, University of Minnesota - Twin Cities, 2024
  • Minor in Computer Science
• B.S. in Physics, University of Minnesota - Twin Cities, 2016

Skills

• Core Skills
  • Machine Learning, Deep Learning, Contrastive Learning, Scientific Machine Learning (Physics/Biology-Informed), Computer Vision, Anomaly Detection, Optimization, Density Functional Theory (DFT), Exploratory Data Analysis, Design of Experiments, MLOps, Signal Processing
• Software
  • Python (PyTorch, Scikit-Learn, Numpy, Pandas, Scipy, Matplotlib, Seaborn, Pymatgen), MatLab, Git/Github, LaTeX, Datalad, Docker, Bash/Unix, Google Cloud Platform (Vertex AI)
• Hardware
  • Fourier Transform Infrared Spectroscopy (FTIR)

Work experience

• Machine Learning Engineer
  Oncodea, October 2024 - Current
  • Develop AI-driven diagnostic solutions to make early-stage cancer screening accessible to the masses.
    
    
• Research Assistant
  University of Minnesota - Twin Cities, April 2015 - November 2024
  • Introduced a Learning Using Privileged Information (LUPI) framework for materials-property prediction, training on costly DFT data but predicting from cheap structural descriptors, achieving 32.5% higher accuracy than standard supervised baselines across multiple datasets.
  • Simulated ~1,000 van der Waals TMD heterostructure candidates via hybrid HSE-functional VASP-DFT, then trained an ML model on monolayer-pair descriptors to predict the heterostructure’s band alignment, achieving 88% test accuracy.
  • Learned an unsupervised SOM representation of 56K materials from Materials Project Database that partitioned a highly non-uniform descriptor space into several local regions, improving downstream classification accuracy by 38% over a global baseline.
  • Formulated a theoretical framework based on Statistical Learning Theory to explain the double descent phenomenon, providing theoretical insights on model complexity in both under- and over-parameterized regimes.
  • Developed a biologically-constrained deep learning model that improves robustness to occlusion, resulting in a 32% decrease in 3D pose error when estimating 3D hand pose from 2D images.
  • Investigated analytical reasoning capabilities of Large Language Models (LLMs) through prompt engineering and proposed strategies for integrating LLM tools into mathematics and machine learning education.
  • Facilitated interdisciplinary technical collaboration between machine learning and materials science experts across multiple research projects.
    
    
• Research Intern
  Astrin Biosciences, May 2023 - August 2023
  • Optimized a deep learning-based real-time cancer cell detection model from holographic images by incorporating physics constraints based on cell morphology and optics, achieving a 10x improvement in specificity while maintaining high sensitivity.
  • Designed experiments to uncover crucial cell features and integrated these new insights to enhance the deep learning model.
  • Developed and deployed a deep learning model for autofocus in holographic cameras, significantly enhancing cell image resolution and removing the need for manual tuning, thereby increasing experiment workflow efficiency by 30%.
    
    
• Research Intern
  Taiwan Semiconductor Manufacturing Company (TSMC), June 2020 - September 2020
  • Secured a patent and trade secrets for inventing computational techniques that enhance multilayer mirror reflectivity in EUV lithography, potentially doubling manufacturing yield.

Fellowships and Awards

• Data Science Initiative (DSI) Fellowship Award, 2023-2024
• Best Poster Award, Midwest Machine Learning Symposium (MMLS), 2024
• Dean’s List, UMN College of Science and Engineering, 2015-2016

Service and Outreach

• Technical Committee, Nanophotonics of 2D Materials Conference, 2020
• Editor, 2D Materials: Properties and Devices Textbook, 2017

Journal Publications

Lee, S., Lee, E.H., Kwon, Y.-K., Koester, S.J., Avouris, P., Cherkassky, V., Tersoff, J., Low, T. Generalized energy band alignment model for van der Waals heterostructures with a charge spillage dipole. ACS Nano 19, 37749–37757 (2025).

Khaw, M.J., Zorko, N.A., Kennedy, P.R., Bendzick, L.E., Shackelford, M., Selleck, C., Hinderlie, P., Walker, J.T., Soignier, Y., Lyons, R.C., Femino, E.L., Stenger, T., Dasgupta, T., Kotz, L.E., Shetty, M., Phung, S.K., Lee, E.H., Lu, Q., Lim, J., Provenzano, P.P., Fujioka, N., Davis, Z.B., Geller, M.A., Wagner, J.E., MacMillan, M.L., Felices, M., Miller, J.S. Novel trispecific killer engager targeting B7-H3 enhances natural killer cell antitumor activity against head and neck cancer. Journal for Immunotherapy Cancer 13, e011370 (2025).

Cherkassky, V. & Lee, E. H. A perspective on large language models, intelligent machines, and knowledge acquisition. arXiv preprint arXiv:2408.06598 (2024).

Lee, E. H. & Cherkassky, V. Understanding double descent using VC-theoretical framework. IEEE Transactions on Neural Networks and Learning Systems 35, 18838–18847 (2024).

Cherkassky, V. & Lee, E. H. To understand double descent, we need to understand VC theory. Neural Networks 169, 242–256 (2024).

Lee, S., Seo, D., Park, S.H., Izquierdo, N., Lee, E.H., Younas, R., Zhou, G., Palei, M., Hoffman, A.J., Jang, M.S., Hinkle, C.L., Koester, S.J., Low, T. Achieving near-perfect light absorption in atomically thin transition metal dichalcogenides through band nesting. Nature Communications 14, 3889 (2023).

Lee, E. H., Jiang, W., Alsalman, H., Low, T. & Cherkassky, V. Methodological framework for materials discovery using machine learning. Physical Review Materials 6, 043802 (2022).

Book Chapters

Cherkassky, V., Lee, E.H.. VC-Theoretical Explanation of Double Descent.. in The Importance of Being Learnable: Essays Dedicated to Alexander Gammerman (eds An Nguyen, K. & Luo, Z.) 157-176, Springer Nature Switzerland, Cham (2026).

Patents

Lee, E.H. and Cheng, W.H. Lithography system and methods. US patent no.: US20220382167A1 (2022).

Posters

Khammanivong, A., Lee, E.H., Khonkhammy, D., Gunnarson, C., Sockalingum, G.D., Bedros, S., Petcavich, R., Pham, D.Q. Abstract LB409: OncodeAi™-Breast CED: A nanoparticle-enhanced IR molecular sensing platform for early breast cancer detection. Cancer Research 85, LB409–LB409 (2025).

Lee, E.H., Lee, S., Feummeler, E., Tadmor, E., Low, T., and Cherkassky, V. Enhancing materials discovery with the LUPI framework: a novel approach to predicting material properties. Knowledge Guided Machine Learning Workshop (2024).

Lee, E.H., Lee, S., Feummeler, E., Tadmor, E., Low, T., and Cherkassky, V. Enhancing materials discovery with the LUPI framework: a novel approach to predicting material properties. Midwest Machine Learning Symposium (2024).

Lee, E.H., Lee, S., Feummeler, E., Tadmor, E., Low, T., and Cherkassky, V. Enhancing materials discovery with the LUPI framework: a novel approach to predicting material properties. 3M Poster Symposium (2024).

Heller, N., Mallery, K., Bristow, N., Travadi, Y., Lee, E.H., Hong, J. Abstract 2304: Detection of early-disseminated cancer cells with deep learning-enabled holographic imaging. Cancer Research 84, 2304–2304 (2024).

Lee, E.H. and Cherkassky, V. VC theoretical explanation of double descent. Institute for Engineering in Medicine Annual Conference (2022).

Lee, E.H. and Cherkassky, V. VC theoretical explanation of double descent. CSE Graduate Students Welcome Session (2022).

Lee, E.H., Grassi, R. and Low, T. Theoretical investigation of phonon-polariton modes in cylindrical hexagonal boron nitride. Undergraduate Research Opportunity Program (2016).

Teaching Assistant

EE 8591: Predictive Learning from Data 2020 / 2022 / 2024

EE 4623: Introduction to Modern Optics 2024

EE 3006: Fundamentals of Electrical Engineering Laboratory 2024

EE 5389: Introduction to Predictive Learning 2021 / 2023

EE 2361: Introduction to Microcontroller 2019

EE 5163: Semiconductor Properties and Devices I 2018

EE 2011: Linear Systems, Circuits, and Electronics 2018