Build Vision Systems That Actually Work
Computer vision isn't about following tutorials anymore. Our programs focus on real implementation challenges you'll face when deploying CV systems in production environments across Taiwan's tech sector.
Explore Programs
Three Learning Tracks That Match Your Goals
We built these paths after working with dozens of engineers transitioning into computer vision. Each track addresses specific career transitions we see regularly in Taiwan's market.
Detection Specialist
Start with object detection fundamentals and move into deployment scenarios. You'll work with YOLOv8, handle edge cases, and optimize for mobile devices — skills that matter when clients need solutions that run on factory floors.
Tracking Engineer
Multi-object tracking feels straightforward until you hit occlusion problems. This track covers DeepSORT, ByteTrack, and practical approaches to maintaining identity across frames in messy real-world footage.
Systems Architect
For engineers ready to design full pipelines. You'll handle camera calibration, coordinate transformations, distributed processing, and the infrastructure decisions that separate hobby projects from production systems.
Learn From Actual Implementation Problems
Most courses teach you to run inference on clean datasets. That's table stakes. What happens when lighting changes throughout the day? When your model needs to run on hardware with 2GB RAM? When stakeholders want results in 48 hours?
Our curriculum comes from consulting projects we've done for manufacturing plants, retail analytics, and security implementations across Taiwan. The messy parts that usually take months to figure out on your own.
Work with datasets that have annotation errors, missing labels, and class imbalances — because that's what you'll encounter in client projects
Debug model failures using techniques we developed while troubleshooting production systems at 3am
Optimize inference speed without sacrificing accuracy — a balance every deployment eventually requires
Programs Starting September 2025
We're accepting applications for our autumn cohort. Both tracks run for 16 weeks with flexible evening schedules designed for working engineers.
Foundation Track
NT$48,000- 16-week structured curriculum
- Weekly live sessions with implementation reviews
- Access to annotated datasets from real projects
- Code review on four major assignments
- Community workspace access in Nantou City
Production Track
NT$72,000- Everything in Foundation Track
- Monthly consulting sessions on your projects
- Priority support via private Discord channel
- Guest sessions with CV engineers from local companies
- Final project deployed to actual production environment
I'd been trying to learn computer vision through YouTube and documentation for about eight months. Made some progress but kept hitting walls when I tried applying things to actual footage from our warehouse cameras. The Production Track gave me frameworks for handling the messy parts — camera calibration, dealing with motion blur, optimizing for our specific hardware constraints. Three months after finishing, I'm leading our quality control vision project.
Dag Vestergaard
Systems Engineer, Logistics Solutions