Many computer vision projects hit a wall after proof-of-concept due to high cost of deployment, complex infrastructure, and maintenance requirements. In this interview, Plainsight CEO Kit Merker discusses why these obstacles persist and how we can rethink the path from prototype to production through reusable, modular, developer-friendly systems.
The conversation centers on OpenFilter, a new open-source project launched May 21, 2025. Designed with modularity at its core, OpenFilter introduces lightweight, reusable components called “Filters” that break down vision tasks, like object detection, OCR, and preprocessing, into configurable building blocks. With a standardized runtime, APIs, and prebuilt examples, it offers a flexible foundation for developers, researchers, and enterprises alike.
If you’ve ever struggled to create a vision AI application, this session offers practical insight into reducing compute costs, simplifying orchestration, and supporting both edge and cloud environments. You’ll walk away with a clearer understanding of why so many POCs fail, and how a new open approach could help you build beyond them.

Top 5 reasons to attend:

Learn why CV projects often fail post-POC
Understand the case for modular, composable architecture
Get insight into OpenFilter’s design and goals
Hear from a leader with real-world platform experience
Take away actionable ideas for making your CV work production-ready