I'm a machine learning practitioner with experience in various data science competitions and real-world projects. I've been fortunate to work in image classification, time series analysis, and medical AI applications, and have had the opportunity to publish research papers and participate in AI competitions.
Recently, I've been exploring Multi-Agent Systems and have had the chance to implement deep learning model uncertainty and data drift monitoring approaches for model lifecycle management in manufacturing environments.
π Competition Participation (Click to expand)
| Competition | Organizer | Project | Tech Stack |
|---|---|---|---|
| Lotte Information & Communications Vision AI | Lotte | Product image classification | Classification |
| Sleep AI Challenge | Seoul National University Hospital | Sleep stage classification based on polysomnography | Classification |
| Competition | Organizer | Project | Tech Stack |
|---|---|---|---|
| Laryngeal Cancer Prediction Model | Korea University Medical Center | AI model for laryngeal cancer prediction from image data | Image Segmentation |
| Konyang Health Datathon 2020 | Konyang University | Classification of benign/malignant breast cancer pathology images | Classification |
| Competition | Organizer | Project | Tech Stack |
|---|---|---|---|
| Konyang Health Datathon 2019 | Konyang University | Classification of eye diseases from fundus images | Classification |
| Competition | Outcome | Rank | Project | Tech Stack |
|---|---|---|---|---|
| University of Liverpool β Ion Switching | π₯ Silver | Top 4% | Predicting open ion channels from simulation data | Time Series Analysis |
| Freesound Audio Tagging 2019 | π₯ Bronze | Top 9% | Audio classification | Audio Classification |
| Bengali.AI Handwritten Grapheme Classification | π₯ Bronze | Top 10% | Bengali handwriting classification | Multi-classification |
| IMet collection 2019 β FGVC6 | π₯ Bronze | Top 13% | Classification of museum artwork cultures and tags | Multi-classification |
| Competition | Organizer | Rank | Project | Tech Stack |
|---|---|---|---|---|
| Naver AI-Rush 2019 | LINE | 5th | CTR prediction based on Japanese LINE article metadata | CTR Prediction |
| Collision Object Detection | KAERI/Dacon | 6th | Predicting collision location, mass, and velocity from accelerometer data | Time Series Analysis |
| Naver AI Hackathon 2018 | Naver | 10th | Image retrieval model development | Image Retrieval |
π Publications (Click to expand)
-
Scientific Reports (Nature)
"Development of machine learning model for diagnostic disease prediction based on laboratory tests" -
MDPI Applied Sciences
"Improved U-Net++ with Patch Split for Micro-Defect Inspection in Silk Screen Printing"
π Industry Learning (Click to expand)
- Contributed to uncertainty-aware model lifecycle management systems
- Monitoring data distribution changes (data drift) in manufacturing environments
- Helping develop mechanisms to quantify model uncertainty for reliability assessment
- Implementing automatic retraining pipelines based on prediction uncertainty
- Assisted in designing real-time model monitoring systems
- Helping develop data drift detection algorithms
- Supporting ML pipeline automation optimized for manufacturing lines
- π€ Multi-Agent Systems
- π§ Medical Image Analysis
- π Time Series Analysis
- ποΈ Computer Vision
- π Defect Inspection Systems
- π Model Uncertainty & Data Drift Monitoring