About
I grew up in Jakarta, Indonesia. Before I chose to pursue Computer Science at BINUS University, I developed a strong interest in physics, statistics, and mathematics. I enjoy learning across disciplines, from medicine and physics to economics. My deepest passion lies in physics, particularly astrophysics, because I am drawn to understanding the unknown aspects of the universe. However, I chose computer science because I believe it allows me to integrate my interests in physics, mathematics, and statistics into developing innovative solutions through Artificial Intelligence and Machine Learning.
I believe that technology, especially AI, has the potential to revolutionize various industries and improve people’s lives. I am motivated to contribute to its growth and development. I enjoy research because it allows me to explore new ideas, push the boundaries of what is possible with technology, experiment with different techniques, learn from existing work, and contribute to the advancement of knowledge in the field.
I am a fast learner, driven by the belief that there are no inherent limits to knowledge, only limits to our willingness to learn and explore. I continuously seek new challenges that foster both personal and professional growth. As a first-generation undergraduate researcher, I self-fund my research through freelance work in software development, teaching, and research methodology consulting for computer vision.
What I do
- AI and machine learning. Where I go deepest. Computer vision with vision transformers, detection and medical imaging, from training runs to models that run on the device.
- Software engineering, MLOps and DevOps. I ship what I build: native apps in Swift, Python services, and the work that takes a trained model into a product, from serving to tracking experiments.
- UI and UX design. I design with Apple's Human Interface Guidelines and Gestalt principles, and I justify each decision. FightRoom was built as a worked example.
- Research. Five peer-reviewed papers in machine learning, plus field, user and design research for the apps I build at Apple Developer Academy.
Projects
- FightRoom. A native iOS 26 app for finding martial arts coaches and gyms, built to show Apple's design guidelines applied in real Swift code.
- 67Cam. A macOS camera studio that turns a webcam into a system-wide virtual camera, with gesture-triggered overlays any app can read.
- DALX. Chest radiograph measurements computed entirely on the Mac, and an honest null when a measurement can't be made.
- OccamViT. A compact vision backbone designed by subtraction, trained from scratch on ImageNet-1K on a single GPU.
- Diabetic Retinopathy Detection Using CNN. Built a Tkinter-based application that analyzes retinal images using a CNN to detect diabetic retinopathy with 88% accuracy on unseen data.
- Facial Recognition for Airport Security. Designed and implemented a real-time facial recognition system using CNN and webcam integration to streamline ticket exchange and enhance airport security.
- Skin Type Detection Using CNN. Developed a real-time skin type detection application that leverages CNN and webcam input to provide immediate analytical feedback.
- Peer Tutoring & Academic Lecture Series. Organized and led free tutoring sessions in Statistics, Discrete Mathematics, and Linear Algebra for 15–20 classmates, along with a bootcamp (IDR 50K) covering Algorithms & Programming, Statistics, and Linear Algebra.
- Deep Learning research on Computer Vision. Conducting research on ensemble distillation and mutual learning for computer vision models.
- Aria Music Player with Vocal/Instrumental Separation. Developed a full-stack music streaming platform featuring real-time vocal and instrumental track separation.
- CS Learning Platform. Developed an interactive education platform with quiz creation functionality, secure authentication, and role-based access control for administrators, lecturers, and students.
- EchoVision. Developed an AI-powered waste classification system with real-time webcam analysis and GradCAM visualization to help users properly sort 12 categories of recyclable materials.
- Factory Bearing Monitoring System. Built a professional industrial IoT dashboard for real-time bearing fault monitoring, featuring animated factory visualization, predictive maintenance, and interactive model analysis reporting.
- Lord Card Shop. Built a full-featured collectible card e-commerce web application with modern UI/UX, secure authentication, transaction management, and responsive design.
Experience
- Apple Developer Academy Scholar, Apple Developer Academy @ BINUS, Feb 2026 – Present
- Undergraduate Research Assistant, BINUS University, Jan 2025 – Dec 2025
- Application Developer Intern, Otoritas Jasa Keuangan (OJK), Jul 2024 – Aug 2024
- General Secretary & Board Member, Bagi Dunia (NGO), Feb 2024 – Present
- Peer Tutor & Educational Content Creator, Independent, Nov 2023 – Dec 2025
Papers
- Enhancing Bone Fracture Detection in X-ray Images Using Faster R-CNN with EfficientNetV2. IEEE ISRITI 2025, 2025.
- CALM: Calibrated Adaptive Learning via Mutual-Ensemble Fusion. Procedia Computer Science (ICCSCI 2025), 2025. DOI
- Analytical Analysis of Cryptocurrency Regulation and Adoption: A Machine Learning-Driven Ablation Study. Procedia Computer Science (ICCSCI 2025), 2025. DOI
- Tiny vs. Tinier: Baseline ViT-Tiny vs. Ensemble-Distilled Student on Imbalanced Fracture Detection. Procedia Computer Science (ICCSCI 2025), 2025. DOI
- Ablation Study: Calibrated Adaptive Learning Ensemble Methodology. Procedia Computer Science (ICCSCI 2025), 2025. DOI
Certificates
- FreeCodeCamp. Scientific Computing with Python, Responsive Web Design Certification, JavaScript Algorithms and Data Structures, Legacy JavaScript Algorithms and Data Structures, Data Analysis with Python, Machine Learning with Python, College Algebra with Python
- Kaggle. Computer Vision, Intro to Machine Learning, Intermediate Machine Learning, Machine Learning Explainability, Intro to Deep Learning, Data Visualization, Data Cleaning, Intro to AI Ethics
- NVIDIA. Building LLM Applications With Prompt Engineering
