Yechan Hong 010-2259-1479 yechanhong@gmail.com github.com/ychnh
Education ---
2016-2019 M.S. Applied Mathematics, Computer Science: University of Missouri (Columbia, Missouri)
GPA: 3.756
Applied Mathematics Supervisor: Jan Segert
Computer Science Supervisor: Jianlin Cheng
2011-2014 B.A. Mathematics, Computer Science: DePauw University (Greencastle, Indiana)
GPA: 3.89
summa cum laude
Professional and Research Experience ---
1.2021-Present AI Researcher: Pharmcadd (Busan, KR)
Protein Structure Prediction
Predict 3D protein from 1D protein sequence by leveraging multiple architectures: LSTM, CNN, Attention and ESM msa embedding.
Experiment and benchmark various loss functions using distance map, dihedral angle maps, frame-aligned point errors.
Protein Contact Map Prediction
Predict protein contact map from 1D protein sequence
Utilize predicted contact map and MODELLER-based tools to predict 3D protein structure.
Filter, process, and standardize a large dataset of 125,000 proteins in various formats.
6.2019-12.2021 AI Researcher: REALTIMETECH (KR)
Satellite Image Segmentation and Classification
Research, implement and benchmark papers: Deeplabv3+, Deep Aggregate Net, Fully Dense Net, U-Net.
Analyzed and implemented a prototype that won a government contract of 1.3 million USD.
Semiconductor-robot monitoring
Research and prototype robot pose tracking algorithm for SK Hynix.
Automatically extract key features from a complex semiconductor process video: robot position, position of nozzles attached to the robot, and the spray/no-spray states of the nozzles.
Overcome techinical challenges by leveraging GPU processing
8.2016-6.2019 Researcher/Teaching Assistant: University of Missouri (US)
Masters Thesis: PRO3DCNN
Research and implement CNN for classifying protein structures into SCOPe folds.
Implement data feature extraction using Persistent Homology that could characterize alpha helices and beta sheets of proteins.
Protein Folding, Protein Sequence Alignment, Nuclei Segmentation
Study and develop bioinformatics algorithms, utilizing key principles: maximal likelihood estimation, simulated Annealing
Teaching Assistant
Lectured, graded, and monitored exams for MATH1100 College Algebra and MATH1400 Calculus for Social and Natural Sciences.
7.2014-2.2016 Research & Development: Epic Systems (US)
Manage and maintain MyChart, a patient medical information software on iOS and Android.
Design, develop, and launch new MyChart software for the Apple Watch.
Awards and Honors
Prototype and proposal for satellite image segmentation and classification won a government contract of 1.3 million USD, Dec 2019
MyChart Apple Watch project show-cased during meeting of 4,000 employees, May 2015
Wylie Condit Science Scholarship, April 2014
Technical Skills
Programming: Python, Pytorch, Tensorflow, OpenCV, Distributed Training