Computer Vision Development Program
Build practical skills in AI and computer vision through hands-on projects and real-world applications
Our program starts in September 2025 and runs for nine months. You'll work with actual datasets, train neural networks, and build systems that solve tangible problems. No magic promises—just focused learning with experienced instructors who've built production systems.
Request Program DetailsLearning Path Structure
We've organized the curriculum into six modules that build on each other. Each one takes about six weeks and includes both theory and practical work.
Python & Data Foundations
Start with Python programming fundamentals and data manipulation. You'll get comfortable with NumPy, pandas, and basic algorithms before moving into vision-specific work.
- Python syntax and data structures
- Working with arrays and matrices
- Data cleaning and preprocessing
- Version control with Git
Image Processing Basics
Learn how computers interpret images. We cover color spaces, filters, edge detection, and basic transformations using OpenCV and PIL.
- Image loading and manipulation
- Filters and convolutions
- Feature detection methods
- Image augmentation techniques
Neural Networks Introduction
Build your first neural networks from scratch, then move to frameworks like PyTorch. You'll understand backpropagation and gradient descent through hands-on coding.
- Network architecture design
- Training loops and optimization
- Loss functions and metrics
- Debugging model performance
Convolutional Networks
Deep dive into CNNs—the backbone of modern computer vision. You'll work with established architectures and learn when to use different approaches.
- CNN layer types and design
- Transfer learning strategies
- Model fine-tuning techniques
- Performance optimization
Object Detection & Segmentation
Move beyond classification to detect and segment objects in images. Work with YOLO, Mask R-CNN, and similar architectures on practical datasets.
- Detection pipeline design
- Bounding box prediction
- Instance segmentation
- Real-time processing considerations
Deployment & Production
Learn to deploy models in real environments. Cover API design, model optimization, monitoring, and handling edge cases in production systems.
- Model serving and APIs
- Performance monitoring
- Edge device deployment
- System integration patterns
How the Program Works
We know people have different schedules and learning preferences. That's why we've built flexibility into the structure without sacrificing depth.
Live Sessions
Two evening sessions per week (7-9 PM Taiwan time) with instructors. These cover new concepts, live coding demos, and Q&A. Sessions are recorded if you need to catch up.
Project Work
Most learning happens through projects you complete at your own pace. Each module has 2-3 assignments that mirror real development tasks. You'll get code reviews and feedback from instructors.
Study Groups
Optional small group sessions organized by participants. Many students find these helpful for working through challenges together and sharing different approaches.
One-on-One Support
Book office hours with instructors when you're stuck. Most students use this 3-4 times during the program for specific technical questions or project guidance.
Support Throughout
Active Discussion Forum
Get help from both instructors and fellow students. Most questions get answered within a few hours. The community sticks around after graduation too.
Resource Library
Access curated papers, tutorials, and code examples. We add new resources based on student questions and current developments in the field.
GPU Access
Cloud computing credits included for training models. No need to own expensive hardware—you'll have what you need to complete all assignments.
Career Guidance
Optional sessions on portfolio building and technical interviews. We share what hiring managers look for based on our experience, but we don't promise job placements.
Linnea Bergström
Manufacturing Quality Inspector
The Challenge
I worked in quality control checking products manually. It was repetitive and I kept thinking there had to be a better way. I'd heard about computer vision but had zero programming background. Starting felt overwhelming.
Learning Process
The first month was rough—Python syntax didn't click right away. But the small group sessions helped. Once we got to image processing in month two, things started making sense because I could see how it applied to my work. By module four, I was building defect detection prototypes.
Current Work
I'm now developing vision systems for the same company I worked at before. They created a new position after I showed them a prototype I built during the program. Still learning every day, but I can actually build things now.
Matteo Vidal
Web Developer
The Challenge
I'd been doing web development for years but wanted to move into AI. Self-study wasn't working—too many scattered resources and no feedback on my code.
What Helped
Having structured projects with code reviews made the difference. The instructors caught mistakes I would've kept making. The deployment module was especially valuable.
Now
Working on a healthcare imaging project. The fundamentals from the program gave me enough to keep learning on the job.
Saskia van der Berg
Research Assistant
The Situation
I was doing research that involved analyzing microscope images. Doing it manually was taking forever. I needed to automate the process but didn't know where to start with machine learning.
Progress
The program gave me the tools to build a custom segmentation model for our specific images. It's not perfect, but it's cut our analysis time by about 60% and I keep improving it.
Freya Johansson
Graphic Designer
Why I Joined
I wanted to understand the AI tools showing up in creative software. Started curious about the technology, ended up fascinated by the technical side.
Current Path
Still doing design work, but now I also consult with clients on AI integration. The technical knowledge helps me bridge the gap between designers and developers. It opened up a niche I didn't know existed.
September 2025 Cohort
Applications open in May 2025. The program costs NT,000 for the full nine months, which includes all materials, GPU credits, and instructor access. We accept 24 students per cohort to keep group sizes manageable.