AI-based OCR Solution, TwinReader (2023. 02 ~ 2024. 09)
- Developed a Kubernetes-based Triton Inference Server for model deployment.
- Resolved source and version control challenges during Triton Inference Server deployment by improving the architecture using GitLab CI and Kubernetes.
- Fixed critical flaws in the Triton Inference Server. (GPU Memory reduced by 47.9%)
- Enhanced inference time of the text detection model using TensorRT. (inference time decrease by 87.31%)

Python Library for Pre/Post-Processing, Visualization, and Model Backend (2023. 10 ~ 2024. 09)
- Maintained the integrity of the library through type hints and PyTest.
- Standardized the preprocessing of various document images and visualization of model inference results.
- Improved the post-processing speed of the text detection model through algorithm enhancements using Cython. (inference time decreased by 74.12%)

Machine Learning Research Engineer

기계공학과

Hyogeun Oh, et al. Classification and Redundancy Quantitative Evaluation Algorithm for Highly Efficient Fault Diagnosis of Rotary Machines in Roll-to-Roll systems,  Measurement (JCR 17.19%), 2023 

Hyogeun Oh, et al. Feature Selection Algorithm Based on Density and Distance for Fault Diagnosis Applied to a Roll-to-Roll Manufacturing System, Journal of Computational Design and Engineering (JCR 10.87%), 2022

오효근

python

머신 러닝

딥러닝

Github

PyTorch