施孟宏 Hiro
軟體研發工程師
台北
熱衷於新科技和技術探索的軟體工程師
擅長 Python 與 2D/3D 深度學習領域的相關技術
並具備敏銳的觀察力和解決問題的能力
工作與求學期間，參與多個新創或AI落地的專案
期待能在充滿挑戰和創新的工作環境與同事們一起學習和成長
專案的介紹可以來我的作品集網站看看
Python
- Deep Learning
- PyTorch
PyTorch Lightning
ultralytics/yolo
Image Processing
OpenCV
scikit-image
PyVista
Data Analysis
Numpy
Pandas
Matplotlib
TensorBoard
GUI
PyQt
開發環境/工具
- OS
- Windows
Ubuntu
IDE
Jupyter
PyCharm
VScode
Tool
Docker
Git
Edge computing
Jetson Nano
NAS
ONNX
PyTorch Mobile
TensorRT
專業課程
- 影像處理
- 計算機視覺
- 類神經網路
- 多媒體概論
- 機器學習
- Python資料分析
- JSP網路程式設計
工作經歷
軟體研發工程師 • 醫療儀器新創
Jun 2023 - Feb 2024
建構深度學習軟硬體環境, 2D/3D 醫療影像資料標註, 3D 醫療影像辨識演算法開發
軟體研發工程師 • QNAP 威聯通科技
Aug 2022 - Dec 2022
NAS 應用程式開發, OCR 框架性能效能驗證, Library 包裝, API 設計
學歷
元智大學 電機工程所
學士 / 碩士 五年一貫
2016 - 2021
專案
Kaggle 競賽銀牌:
APTOS 2019 Blindness Detection
使用 EfficientNet
搭配多種深度神經網路的訓練技巧
建立一套全球排名 Top 1% 準確度的
糖尿病視網膜病變分類網路
施孟宏 Hiro
Software Engineer
Taipei
A software engineer, new tech lover
Thinking with observation and logic
Teamwork to solve projects with Python and 2D/3D deep learning
Participate in multiple startup or new technology projects
More project descriptions are shown on my portfolio website
Python
- Deep Learning
- PyTorch
PyTorch Lightning
ultralytics/yolo
Image Processing
OpenCV
scikit-image
PyVista
Data Analysis
Numpy
Pandas
Matplotlib
TensorBoard
GUI
PyQt
Dev Env/Tools
- OS
- Windows
Ubuntu
IDE
Jupyter
PyCharm
VScode
Tool
Docker
Git
Edge computing
Jetson Nano
NAS
ONNX
PyTorch Mobile
TensorRT
Courses
- Image Processing
- Computer Vision
- Neural Networks
- Introduction to Multimedia
- Machine Learning
- Python Data Analysis
- JSP Applications
Work Experience
Software Engineer • Medical Equipment Starup
Jun 2023 - Feb 2024
2D/3D medical image data processing, 3D medical image recognition algorithm development
Software Engineer • QNAP Systems, Inc.
Aug 2022 - Dec 2022
NAS Application Dev, OCR framework performance validation, Library packaging, API design
Education
Master of Electrical Engineering, Yuan Ze University
Accelerated Five-Year BA/MA Program
2016 - 2021
Projects
Silver medal in Kaggle competition
APTOS 2019 Blindness Detection
Use EfficientNet with tricks
Top 1% Accuracy in the World
Diabetic Retinopathy Classification Network
Sphygmomanometer Digits
Recognition System
Through the camera lens of a mobile device
Capture clear images from the screens of two target sphygmomanometer
Detect the digits on the computing server
Shorten the procedure when using a traditional sphygmomanometer to record in the cloud
Deinterleaving and Classification of Multiple PRF Signals
Designing radar signal encoder and neural network
Identify and classify multiple radar simulation signals
