Ming-Jen Chang Email & Phone Number

Chicago, Illinois, United States

Automatically Quote Shipping Price – Python, HTML • Developed real-time price quote tool using zip code mileage and weekly updates from market rates via a web crawler.• Automated per-mile rate updates, ensuring pricing stays competitive and accurately reflects current market conditions.• Enabled clients to estimate accurate shipping costs instantly, boosting convenience and reliability in price transparency.

Chicago, Illinois, United States

Scalable Similarity Image Search Engine• Utilized the ResNet-50 deep learning model to extract feature embeddings, enabling high-efficiency image similarity comparisons.• Implemented a top-k retrieval system for the ImageNet dataset, providing scalable solutions for image search functionality.• Optimized model performance for quick retrieval, ensuring rapid and relevant matches for large-scale image datasets.Purchase Website Connected with SQL Server• Created a web-based trade system integrating front-end and back-end using SQL Database and C# for smooth transactions.• Designed an intuitive user interface connected to a robust SQL backend, optimizing seamless trading and database management.

Airlines Analytics Based on Data Visualization• Analyzed flight delays for Delta and Alaska Airlines using Tableau, revealing frequent delay causes under different scenarios.• Visualized airline performance metrics, helping to identify delay trends and optimize strategies for on-time performance.Credit Default Forecasting• Classified credit defaults and predicted new customer classes with machine learning, achieving accuracy rates up to 82%.• Developed a predictive model utilizing diverse data processing techniques, boosting forecasting precision in credit risk assessment. Analyzing Co-Purchasing Items on Amazon• Analyzed and clustered Amazon co-purchasing data with R to uncover product relationship patterns in massive datasets.• Employed clustering techniques to reveal insights into co-purchase behavior, helping identify related products effectively.Predictive and Analyzing the Price of Diamonds• Analyzed and predicted the price with the relevant parameters via data visualization and machine learning algorithm

Chicago, Illinois, United States

Word Predictive System• Predicted the next k words of the sentence while typing based on the LSTM model

Taipei City, Taiwan

Indoor Positioning System• Developed an indoor positioning system utilizing Bluetooth signals and proprietary algorithms to track and navigate within enclosed spaces.• Integrated real-time location tracking, enhancing the system’s accuracy for indoor navigation across complex environments.Automatically Testing on Android Compatibility Test Suite• Automated over 10 manual testing processes in the Android Compatibility Test Suite, saving 60% of the testing time and boosting efficiency.• Increased test coverage and reliability by automating compatibility tests across various Android devices and versions.Customized User Interface with Android System• Customized the Android system’s volume interface and main screen volume bar, improving user experience with visual feedback.• Collaborated with clients and colleagues to define project requirements and ensure timely progress, ensuring high-quality results.

Taipei City, Taiwan

Data Management Server• Designed and implemented an internal data management server to efficiently organize and label image data from department-wide projects.• Improved data handling capabilities, streamlining workflows and ensuring faster processing for large datasets across the team.

Taipei City, Taiwan

Artificial Intelligence Software DevelopmentFacial Expression & Emotion Recognition• Utilized Open VINO and YOLO-v4 to recognize facial expressions and emotions, using 64 landmark points for precise detection.• Employed advanced computer vision algorithms to accurately detect various emotional states and facial movements.Vehicle Detection & Recognition System• Trained deep learning models for vehicle recognition, running efficiently on ARM-based devices like Raspberry Pi and Nvidia Jetson.• Enhanced recognition speed and accuracy, enabling deployment in resource-constrained environments.Automatically Labeling & Quality Checking of IC Chip Images• Integrated a trained deep learning model with a server to automate image labeling, reducing manual labeling by 90%.• Enhanced image quality control by automatically checking and tagging image data for improved consistency.

Taipei City, Taiwan

Thesis: A vision-based human action recognition system for moving cameras through deep learning • Recognized human actions effectively when the camera of a robot is moving toward the target person from various directions• Utilized multiple types of data as input information—RGB sequences, Deep sequences, and optical flow sequences— respectively input into three modified 3D CNNs, these recognition results are integrated to output the final recognition result of human actions • Improved 1.5% of the accuracy through modification and ultimately obtained a total accuracy of 96.4%Published onProceedings of the 2019 2nd International Conference on Signal Processing and Machine Learning (SPML), pp. 85–91. 2019.Chang, M.J., Hsieh, J.T., Fang, C.Y., and Chen, S.W.

Master Of Science - Ms, Business Analytics

• Proficiency in complex data structures is essential for effective analysis, ensuring organized and accessible information.

Master Of Science - Ms, Computer Science And Information Engineering

• Proficiency in deep learning involves understanding and expertise in advanced neural network architectures and algorithms, enabling.