Bridging Tech and Practicality for Real-World Impact

At the Group Engineering Centre (GEC), Nah Wu, a Lead Software Engineer in Video Analytics, bridges the gap between advanced technologies and functional, user-friendly solutions.

Whether it’s streamlining machine learning processes for DroNet monitoring or refining video analytics for surveillance, Nah Wu’s work ensures that our common technology modules don’t just remain concepts – they become valuable tools that accelerates our Group’s innovation and success.

Q: Tell us what you do at work.

Our team develops common technology modules that can be adapted and deployed across diverse applications, allowing us to scale solutions quickly. I focus on identifying use cases for these modules, ensuring we can productise them with other tech stacks in a cost-effective, scalable way. The aim is to make these technologies as effective and easy to use as possible, so they’re valuable to customers right out of the box.

Q: What are some of the common technology modules you’ve been working on?^[1]

Among our common technology modules, I’ve worked on VisionX, a generic text search capability for videos and streams; VisionSuite, a one-stop platform that leverages different video analytics engines for AI-driven surveillance; and REVIDA (REusable VIdeo Data Analytics), which streamlines machine learning operations with automation and built-in utilities.

Q: What are some challenges you face in making our modules market-ready?

A big challenge is balancing advanced technologies with user experience. It’s not enough for a product to work in a controlled lab – it must perform in real-world conditions. For example, we tested VisionX with varying lighting and environmental conditions to ensure reliability and ease of use. This process requires rigorous testing and continuous refinement. 

Q: How do you ensure these solutions meet industry needs? 

We work closely with teams across the Group to tailor these modules for specific needs. When we developed AGIL® Vision, we refined product requirements with the Digital Systems team to develop Vision AI modules that address targeted customer needs. Their feedback helped us create a more versatile tool that’s adaptable to different surveillance environments. 

Similarly, we’re collaborating with the Smart Mobility Road team to create a Vision AI architecture capable of processing real-time video streams from thousands of cameras for traffic monitoring. Such cross-functional efforts ensure our innovations are both technically sound and practically deployable.

Q: What is your approach to integrating new technologies such as GenAI? 

GenAI is not about replacing existing technologies; it’s about addressing challenges that weren’t solvable before. Unlike traditional AI, it requires new hardware considerations and thorough evaluations of its accuracy for real-world applications. The key is finding the right balance between cost, precision, and scalability, ensuring that GenAI complements our existing tools effectively.

 ^[1] These are working names of our common technology modules, not actual product names.