Computer Vision Training Programs

Real projects. Hands-on guidance. Programs that start in autumn 2025. We built these courses around what companies actually need when they hire vision developers—not just theory dumps.

Get Program Details

Students working on computer vision projects in collaborative lab environment

Foundation Track

Start With Working Code

Forget passive lectures. You'll write your first object detection script in week one. By week three, you're building a real-time tracking system.

Our foundation track runs sixteen weeks starting October 2025. You work with OpenCV, learn how cameras see the world, and build actual tools someone might use.

Small groups mean you get feedback on your code before bad habits stick. And the projects? They come from real problems we've solved for clients in Taiwan's manufacturing and retail sectors.

Who Teaches These Programs

People who've shipped production vision systems and remember what it's like to be stuck on a problem at 2 AM.

Oleksiy Shevtsov

Lead Instructor

Spent eight years building vision systems for warehouse automation. Still codes every day and actually enjoys debugging your camera calibration issues.

Linnéa Bergström

ML Systems Specialist

Builds training pipelines for defect detection. Knows why your model works in the lab but fails on the factory floor—and how to fix it.

Aisling O'Connor

Applications Engineer

Deployed retail analytics systems across fifteen locations. Teaches the unsexy parts—like dealing with terrible lighting and handling edge cases.

Common Questions About Our Programs

Most people ask these before enrolling. Click any question to see a real answer—not marketing speak.

What programming background do I need?

You should be comfortable writing Python without constantly checking documentation. Know what a dictionary is, how loops work, basic file handling. If you've built something—anything—with Python before, you're probably ready. We don't expect ML experience. That's what you're here to learn.

How much time should I expect to spend?

Classes run twice weekly in the evenings—three hours each session. Then budget another eight to twelve hours per week for projects and practice. Some weeks lighter, some heavier depending on what you're building. It's designed for people with day jobs, but it's not casual learning. You'll actually work.

Do you provide hardware or computing resources?

Lab sessions include access to our cameras and testing rigs. For training models, we provide cloud GPU credits so you're not burning out your laptop. You'll need a decent computer for development work—nothing crazy, but a five-year-old machine might struggle. We can discuss specifics if you're unsure what you have will work.

What happens after I complete the foundation track?

Some people take what they learned and apply it at their current job. Others continue with our advanced track starting early 2026—that's where we dig into custom architectures and deployment optimization. A few join project teams we run for local companies. There's no single path, and we don't make promises about outcomes. But you'll have real skills and a portfolio showing you can build things.

How The Sixteen Weeks Actually Work

Here's what you build and when. This isn't a rigid schedule—we adjust based on how the group progresses.

Weeks 1-4

Image Processing Fundamentals

Get comfortable reading and manipulating images programmatically. Build filters, detect edges, understand what pixels actually are beyond tiny colored squares.

Project: Create a tool that finds manufacturing defects in sample images. Basic but satisfying when it works.

Weeks 5-8

Object Detection Systems

Move from still images to video streams. Learn how existing models work and when to use which approach. Deal with the frustrating reality of inconsistent lighting and camera angles.

Project: Build a counting system for retail environments. Track objects moving through frame without double-counting.

Weeks 9-12

Training Custom Models

Learn what happens under the hood when you train a neural network. Collect and annotate your own dataset. Discover why more data doesn't always mean better results.

Project: Train a classifier for a specific use case. Measure its performance honestly and improve it systematically.

Weeks 13-16

Deployment And Real Conditions

Take your working prototype and make it production-ready. Handle edge cases, manage errors, optimize for actual hardware constraints. This is where theory meets messy reality.

Final project: Something that could actually run in a business environment. You present it to the group and defend your technical choices.

Applications Open June 2025

We're keeping groups small—probably around twelve people per cohort. If this sounds like something you want to do, reach out early so we can talk through whether it's a good fit.

Start A Conversation