This personality profile is part of a series of stories about Xponents who have been with HTX since the very beginning. Their contributions have been pivotal to the growth of the agency, which celebrates its 5th birthday on 1 Dec 2024.
Professor Ng Gee Wah is understandably taciturn when it comes to discussing his role as a director and Distinguished Member of Technical staff (DMTS) at DSO National Laboratories. The blueprinting of Singapore’s national security technologies — which falls under the research institute’s remit — is for obvious reasons, veiled in secrecy.
He is, however, more forthcoming about his work at HTX, which you can scarcely swing a stick at without hitting in the public domain.
“It’s very satisfying to see technologies we’ve developed being used by Home Team Departments (HTDs) such as the Central Narcotics Bureau of Singapore (CNB) to identify potential areas containing drugs, or helping police officers to be more productive in fighting scams,” said Gee Wah, who is the director of HTX’s Q Team Centre of Expertise (CoE).
“I myself receive phishing messages from time to time, so it hits close to home,” he added.
His department at HTX is underpinned by the mission of spawning prototypes at a rapid pace to “solve wicked problems” — that is, unconventional challenges that may elude external vendors. Its multi-disciplinary projects span the areas of cybersecurity, robotics and unmanned systems, and Artificial Intelligence (AI) and data analytics. The lattermost field, a buzzy tech trend crowed about by e-commerce retailers and venture capitalists alike, is a mainstay of his career.
An engineer by training, Gee Wah holds a PhD in Electronics and Electrical Engineering with a specialisation in artificial neural networks, which are software algorithms designed to mimic how the human brain works to recognise patterns, gather data insights, and make decisions or predictions. He latched onto his research topic in the 90s, back when generative AI was still an arcane concept largely hashed over by those in the STEM fraternity. Those discussions, though, hummed with promise.
“Even in the late 80s, we were already talking about artificial neural networks as the next generation programming language, which could generate output without having to do coding,” he recalled. “There was a lot of hype then, so I joined in out of curiosity.”
New frontiers in problem solving
Gee Wah’s prosaic account of his foray into AI and data analytics belies his passion for the subject. He thanks DSO for giving him the opportunity in his early career journey. This is evident in the way he expounds on AI’s potential to drive broader applications, painstakingly drawing multiple analogies to demonstrate its limitations and possibilities.
During our conversation, he branched into explanations of the latest OpenAI o1 large language model’s (LLM’s) capabilities and also its shortfalls in complex tasks such as mathematical reasoning. He pointed out that malicious actors may exploit advances in AI technology to carry out scams and other harmful activities, and that enabling the Home Team to stay several paces ahead of such malicious actors is a “cat and mouse game” which warrants the deployment of a suite of tools.
Helping to build such capabilities to safeguard homeland security is an interesting change for the veteran engineer, who served at DSO for more than two decades. There, he was instrumental in developing AI algorithms for various defence applications.
“When I first started out at DSO, I was looking at national security; now I am looking into internal or homeland security. This makes my working life complete as I have a well-rounded understanding of defence strategies and domestic security operations,” he said.
Some five years ago, he received a call about helping to spearhead a new team at HTX, to which he was receptive. At the start, working with a team with uniformed officers posed some challenges as they weren’t necessarily technically inclined, he noted. But he rolled with the punches nonetheless, nurturing his teammates and playing to their strengths.
“Our team has grown; we’ve mentored and trained staff who are good, self-motivated people, allowing us to address many issues in HTDs despite our small size,” he shared.
He also noted the Home Team’s waning reliance on vendors to roll out new products that are strategic to Home Team over the past five years, thanks to his team and HTX COEs’ strides in bolstering its engineering and digital capabilities.
Failure is no obstacle
On his part, Gee Wah said he eschews hierarchical structures that can hobble innovation.
“I like things to be done fast, particularly in the space of AI where technologies do not wait for you. This is why I prefer to keep discussions, processes and execution procedures relatively flat,” he shared.
Among the most transformative projects delivered by the Q Team, is QCaption, a state-of-the-art multimodal generative AI tool that helps police officers annotate crime scene evidence, generate incident reports and screen videos. It achieves this by generating text descriptions for images and video clips in real time.
“Then, there wasn’t a commercially available product that could perform the same functions; hence we developed the technology and successfully trialled it. We’re currently transitioning to full production,” he shared.
Gee Wah posing with the many gadgets that Q Team has developed over the years. (Photo: HTX)
But he doesn’t just celebrate the initiatives that have taken off. In fact, HTX recognises projects that do not pass muster through its Undaunted Award. Inaugurated in 2023, the awards are premised on driving improvement by dint of embracing failure.
In 2024, Gee Wah’s team won an Undaunted Award for its Q-Crowd Counter, which monitors crowd sizes at large-scale events using live camera feeds from drones. When trialled at dance music festival ZoukOut 2022, the system demonstrated diminished accuracy at night, due to ambient conditions and flashing lights.
Gee Wah regards the award as impetus to continue iterating, as “the so-called failed projects have been a learning journey for the team.”
This persevering attitude may explain his love of running. Outside of work, the father-of-three observes a decidedly predictable routine which revolves around jogs with his son, church on weekends and tutoring his kids. One of them has special needs, thus requiring a greater degree of care.
“She’s reached a stage where she’s quite talkative now, but we sometimes can’t understand what she’s saying, so we need to provide her with more attention,” he revealed.
As methodical as this team leader may appear, he betrays an empathetic side that values staff welfare over notching successes. For instance, he said that mothballed projects don’t lower his morale as much as attrition, as technical issues are par for the course.
“When a staff member leaves you to join another company, you’re left wondering if you didn’t manage him properly and caused him to move on,” he reflected.
Thankfully, he added, those are rare occurrences in the engineers he’s managed.
For now, he maintains his pin-sharp focus on “rapid prototyping and transiting to products that are useful for the Home Team — which is the same goal I had when I first came to HTX”.
Importantly, he’s happy that his work allows him to marry his passion with meaningful practicality for the homeland security.
“This is the best way to channel your energy in life. After all, you only have one life on earth,” he mused.