Master of Science (M.S.) in Electrical Engineering and Computer Science (EECS)

• CGPA: 4.04 / 4.30
• Advisor: Professor Hsien-I Lin 林顯易 教授
• Laboratory: Human-Centered Robotic and Automation Laboratory
• Research Interest: Vision-Language Model, Image Processing, Computer Vision, Generative AI
• Research Topic: Detection Vision-Language Model for Robotic Assembly
• Relevant Coursework: Data Mining, Machine Learning for Signal Processing, Pattern Recognition, Deep Learning, Digital Image Processing, Mobile Robots, Applications of Language Processing in Practice, Computer Vision, Practice in Entrepreneurship

Bachelor of Engineering (B.Eng.) in Information Engineering (IT)

• Final CGPA: 3.65 / 4.00 (Cum Laude)
• Thesis Advisor: Prof. Ir. Hanung Adi Nugroho, S.T., M.Eng., Ph.D., IPM, SMIEEE, MACM
• Thesis: Online Lectures Monitoring Software Component Development and Engineering
• Relevant Coursework: Algorithms and Data Structures, Artificial Intelligence, Image Processing and Computer Vision, Natural Language Processing, Big Data and Analytics, Database Engineering, Software Engineering

• Joining Human-Centered Robotics and Automation Laboratory (Hucenrotia Lab), NYCU in the Generative Artificial Intelligence (GAI) team.
• Created and arranged the system architecture for utilizing LINE Bot with ChatGPT in Azure cloud environments.
• Managing the laboratory AI and computing server PC as the main administrator.
• Researching and creating an integrated vision system utilizing Vision-Language Model (VLM) and a function for object detection also object coordinate location.
• Skills: Python, Image Processing, Computer Vision, Deep Learning, Machine Learning, Generative Artificial Intelligence, Natural Language Processing, System Architecture and Integration, Project Management

• Taiwan Experience Education Program (TEEP) in Human-Centered Robotics and Automation (Hucenrotia) Laboratory, NYCU. Joining the Collaborative Autonomous Mobile Robots research group.
• Created a program flow for robotic arm in the autonomous mobile robot using TMFlow for Awarding Ceremony.
• Research about implementing Inverse Kinematics for Isaac Sim simulation.
• Research and ideation about robotic arm end-effector error calculations and end-effector manipulability.
• Skills: Python, Natural Language Processing, TMFlow TM Robot, Isaac Sim NVIDIA, Robotics

• Helped Senior Data Engineer with daily maintenance tasks, urgent tasks and issues, tech backlog, and system monitoring.
• Created Automate Address Domicile Prediction and Correction, an in-house program and algorithm to predict and correct domicile name aliases, using pattern matching technique and dictionary, solving and correcting valid but incorrect domicile name aliases. The accuracy results for correcting city domicile name aliases reached 79,1% and for correcting district name aliases reached 97,6%.
• Created Personal Identifiable Information (PII) Scanner, an in-house algorithm, model, and program to scan, profile, and tag PII from database tables. The created algorithm and model accuracy outperforms Google Sensitive Data Protection [formerly Google Data Loss Prevention] for Indonesian PII data. The accuracy result of Google Data Loss Prevention at the time was around 35-50% meanwhile the created algorithm and model reached 75-91%.
• Research and defining bad data and garbage data within data sources. The definition leads to ideation how to treat those and created a Looker mock-up design of usable implementation in form of table report for data exploration.
• Skills: Python, SQL, Machine Learning, Natural Language Processing, Data Analysis, JupyterLab, BigQuery, Docker, Looker Studio, Airflow, Streamlit

• An empirical study of three systems implementations: fine-tuned Transformer-based with early and late fusion, zero-shot Auto-Regressive (AR) LLM based, and zero-shot Diffusion LLM based.
• All systems follow the same two-stage pipeline then leverages bilingual fusion by combining original and cross-translated tweet and integrates annotator demographic.
• Utilized Python, PyTorch.
• Paper published. Joy Chrissetyo Prajogo, Lung-Hao Lee, and Hsien-I Lin (2025). NYCU-NLP at EXIST 2025: An Empirical Study of Annotator-Aware Two-Stage Pipeline for Sexism Detection in Tweets. In Proceedings of the Working Notes of CLEF 2025 - Conference and Labs of the Evaluation Forum, Madrid, Spain, 9-12 September 2025, vol. 4038, pp. 2119-2132.

• Implemented Dual Dense Upsampling Convolution (DDUC) for road scene semantic segmentation using the Cityscapes dataset, based on ICCEA 2024 research.
• Integrated DDUC into the SegProp framework for semi-supervised segmentation of aerial videos using the RuralScapes dataset.
• Developed a video inference pipeline to generate segmented video outputs from raw input footage.
• Built a module to quantify the Quality of Green Urban areas from semantic segmentation results.
• Utilized Python, PyTorch, OpenCV, and NumPy for model development and evaluation.

• Conducted contrail segmentation on satellite imagery using deep learning models to support climate change mitigation in aviation.
• Compared segmentation architectures including UNet, EfficientNetB7+UNet, and CBAM+UNet, using a 450GB dataset from a Kaggle competition.
• Standardized training with Adam optimizer, 11 epochs, and consistent hyperparameters across all experiments.
• Evaluated model performance using DICE Score, testing both BCE and BCE + Dice loss functions.
• Achieved highest performance with EfficientNetB7 + UNet using BCE loss, improving segmentation accuracy over the baseline.

• Developing the vision system of the Robotic Generative AI 3-Cup and Ball Game project.
• The vision system will utilize YOLOv7 to detect the cups and ball then use SORT algorithm to track the ball and the cups. This detection and tracking will be combined with custom algorithm to track the ball inside the cup.
• The project is showcased in the Intelligent Asia 2024 in Nangang, Taipei, Taiwan on August 21-24, 2024, inside one of the areas within the National Science and Technology Council booth.
• Paper published. N. Khoirurizka, J. C. Prajogo, S. Perkins, C. -H. Kuo and H. -I. Lin, A Case of Cups and a Ball: Utilizing Generative Artificial Intelligence for Human-robotic Collaboration in Task Execution, 2024 International Automatic Control Conference (CACS), Taoyuan, Taiwan, 2024, pp. 1-6, doi: 10.1109/CACS63404.2024.10773301.

• Created a Python program utilizing OpenCV and Numpy.
• The program implemented Corner Detection to get the tracking points and then Optical Flow to track the points with the goal is to calculate the displacement for the shakiness.
• The input is a shaky video, and the output is a stabilized video that shakiness or unstable factor has been removed.

• Created a Python script utilizing OpenCV, Numpy, and Tensorflow.
• The program implemented several image processing techniques and U-Net architecture for segmentation the cancer and tumour areas, then DenseNet201 for feature extraction and classification part to get the features of the segmentation and to classify the cancer or tumour whether it is malignant or benign.
• The metrics used in the project are Region Over Curve (ROC), Area Under Curve (AUC), and Accuracy.
• The input is ultrasound medical images with cancer and tumour cases.

• Created a Python script utilizing Pandas, Numpy, NLTK, Scikit-Learn.
• The script will implement data selection techniques, data processing with 6 techniques or methods, and then classification.
• The metrics used in the project are Accuracy, Precision, Recall, and F1-score.
• The input is Tripadvisor hotel reviews data. The output is a statistical report and analysis.

• Helping the President of NYCU Indonesian Student Association to arrange events.
• Guiding and helping new and current Indonesian students in NYCU with their campus life.

• Leading a student group of 30 members.
• Plan and do 102 programs for village communities in Yogyakarta, Indonesia in 50 days.
• The programs made to empower the local communities have various goal areas, from economic development, environmental, technology, social studies, public health, and others.

• President of Night Login Data Science and Data Analytics Community (NDDC), a division of a student body community for Information Engineering students.
• Creating training curriculum related to Data Science and Data Analytics for the members.
• Give the training to the members.