Artificial intelligence (AI) has gained significant attention in recent times, thanks to the emergence of ChatGPT. But it is seldom noticed that AI has been extensively developed in online advertising over the last two decades.As digital advertising continues to dominate a colossal market share worldwide, Hong Kong stands out as an international city experiencing consistent growth in ad spending within the Asia Pacific region. With an estimated USD 1747 million allocated to digital advertising in 2024, the landscape is ripe with opportunities. To shed light on this dynamic field, we have interviewed Jeff Lao, Senior AI Engineer at HKAI, to ask him his incentive in joining a new start-up company, which focuses on enabling AI on online advertising. His experience in this start-up will offer us some insight on the development of AI in the edge cutting industry of online advertisement.
Why did you choose to join HKAI?
After graduating from the master’s program in Computer Science, specializing in Intelligent Robotics, at the University of Southern California. I returned to Hong Kong to advance my career. And yet, I found limited job opportunities where I could fully utilize my AI knowledge and nurture my intellectual interests for research and high technologies. That’s when I discovered a startup in Hong Kong building its own programmatic advertising products, with the goal of catering the unique needs of local publishers and advertisers through ad serving engine. This idea fascinated me, and I wanted to be one of the pioneers to make the AI engine real in Hong Kong. Since embarking on this journey, I have been working at HKAI for two years already.
Jeff and the HKAI team celebrating the company’s third anniversary
Could you share briefly what is your role as a Senior AI engineer at HKAI?
As a Senior AI engineer at HKAI, my role encompasses various responsibilities. First, I stay abreast of the latest AI research and development. Given the rapid evolving nature of AI, it is crucial for me to understand the latest machine learning models. This ensures that our ad serving product, AIgoAD, remains up-to-date and utilizes the most advanced model for inventory forecast. It enables us to deliver ads based on audience’s interests while keep improving the click-through rate (CTR) of advertisement.
In addition, I work on consolidating vast amounts of data and algorithm designs that enable machines to learn from them automatically. This plays a crucial role for the machines to make intelligent decisions within our AI systems, so as to optimize the accuracy and efficiency of the AI models we create.
Daily work of Jeff and his colleague in training AI at the office
In the last two years at HKAI, which project are you most proud of and how it has benefited the clients?
The project I am most proud of is the creation of the “user look-alike model” in AlgoDATA. This project has greatly helped our clients to reduce the operational cost and computer processing time while delivering accurate recommendation for enterprises to expand their potential customer base.
Previously, our clients had to manually input rules to an external software to generate a list of potential customers. Such software uses statistical approaches to compute the rules, which took approximately 10 days to generate results. However, by leveraging our AI technology, clients can now bypass the rule selection and evaluating process. Instead, they can entrust AI to process consolidated numerical representation of the customers’ interest and minimize the computation resource from a cloud cluster to a laptop. This significantly reduces the processing time, allowing clients to obtain precise results within just a few hours.
What’s the best part working at HKAI?
It must be the culture of collaboration and innovation without hierarchical constraints in HKAI. Unlike large tech companies where product improvement plans can be challenging to implement, at HKAI, we have the autonomy to create the best solutions for our clients. For instance, I once initiated a new project for data consolidation in our database to enhance AI-based inventory forecasting, an idea inspired by an open-source discussion. At HKAI, we can work towards our clients’ best interests without burdensome processes of approval and administration.
Moreover, HKAI values everyone under its employee-friendly policies. We enjoy flexible work arrangements, such as remote work options once a week to save commute time, which is favorable to me as it takes one and a half hour for me to reach the office. Additionally, we have the opportunity to take “personal time-off” to handle our own financial matters before the banks close. We can maintain a healthy work-life balance here in HKAI.
Editor’s note: Whenever there is a delightful aroma of coffee floating through the office, it must be coming from Jeff Lao. He grinds and brews his coffee every morning and after lunch, giving out the freshest flavors. You can find all the necessary coffee-making essentials neatly arranged on his desk. Maybe coffee also helps him to train the AI smarter too?
