Development PracticesDevelopment Practices
Quickie15min
INTERMEDIATE

Building a multi-tier AI system: From Edge AI Cameras to Cloud-based AI

This talk presents Moverse Capture’s markerless 3D motion-capture architecture, spanning edge AI cameras, low-latency local inference, and cloud services. It examines evolving design decisions, failed assumptions, heterogeneous AI integration, Python-to-specialized-backend bridges, local-first recording, and streamlined cloud deployment, offering practical systems-engineering insights from product development.

talk.summaryAiDisclaimer

Nikolaos Zioulis
Nikolaos ZioulisMoverse

talkDetail.whenAndWhere

Friday, November 20, 15:55-16:10
Room 1 - Thessaloniki
talks.description
This talk will explore the design, architecture, and systems engineering behind Moverse Capture, a markerless 3D motion capture platform spanning edge AI cameras, a real-time and low-latency local inference engine, and an accompanying set of cloud services. Following the progression from our earliest architectural decisions to the challenges that shaped the product today, the talk serves as the implementation-focused sequel to a prior scientific presentation. The main theme is the complexity of building a product that embeds AI across every heterogeneous tier: we will discuss the pros and cons of our approach, where initial assumptions failed, the insights we gained, and how we bridge Python AI code with specialized inference backends, local-first live recording, and streamlined cloud service deployment.
motion
architecture
capture
ai
talks.speakers
Nikolaos Zioulis

Nikolaos Zioulis

Moverse

Greece

Nikolaos (Nick) Zioulis is an engineer and scientist working at the intersection of computer vision, machine learning, and computer graphics for over 10 years. As the CTO of Moverse, he architects and leads the development of a AI-driven 3D motion capture system, aligning its roadmap with business goals.
Before joining Moverse, Nick spent over eight years as a Research Associate at CERTH’s Visual Computing Lab, leading R&D on 3D vision. He holds a PhD in Computer Vision from the Universidad Politécnica de Madrid and an MSc in Electrical & Computer Engineering from the Aristotle University of Thessaloniki.