Hunjae Jung and the work behind the click

In March 2020, Hunjae Jung encountered a database company through an Australian engineer visiting South Korea. Snowflake did not yet have a Korean office. What caught Jung’s attention was a practical proposition: storage and computing could be paid for separately. For someone building software that had to keep swallowing other companies’ data, the arrangement had a particular charm. The interesting thing about a new technology was what it might let a small team stop worrying about.

Two years later, he could put numbers to that decision. Traffic had increased fifteenfold since adoption, he explained, while the DevOps team had grown from one person to four. The arithmetic is more revealing than a founder’s ceremonial photograph. Work had multiplied. The people required to keep it moving had increased much less. Somewhere inside those unequal rates was room for the business to grow.

Jung is now co-founder and co-CEO of Airbridge and AB180, the Seoul-based company behind it. His route to the job ran through the technology and the product: he previously held the CTO and CPO roles. To understand the executive, it helps to spend a little time with the machinery.

A destination inside the app

Before AB180, his project work included Curie, a mobile search engine based on deep linking and app indexing, developed in late 2014 and early 2015. Another 2015 project trained a neural network on 80,000 Flickr images to recognize styles and objects. These were different problems, but both involved making information easier to find or classify. They give the later marketing work a technical prehistory. Search, images and app content already occupied his attention before the marketing dashboard became the main stage.

Start in June 2015. Jung was explaining mobile deep links on AB180’s blog. The subject concerned a familiar irritation: a link ought to take you to the thing you wanted. On the web, that might be a particular page. Inside a mobile application, the destination could be a song, a product or another piece of content hidden behind the app’s front door.

His explanation went through the awkward variations. A person might use iOS or Android. The app might already be installed, or a trip to the app store might come first. The desired destination remained the same; the route had to accommodate the traveler. A good link was doing more work than its modest appearance suggested.

AB180’s early business also changed direction. It began with search, moved toward making app content discoverable, then encountered a customer request: could it explain where app visitors had come from? That question led toward advertising attribution. Getting someone to a destination and understanding the journey proved to be neighboring problems.

There is a pleasing economy to that origin. A customer had supplied the next question. Following it meant building something different from the original plan. The click, so small on the screen, had opened a larger engineering job.

The numbers need somewhere to live

Airbridge connects marketing activity with behavior across websites and apps. A dashboard gives the arrangement a civilized face. Underneath, information has to arrive, be processed and become available in time to help someone make a decision. A delayed answer can be technically correct and commercially rather late for dinner.

By 2022, the infrastructure was handling more than two billion events a day. Customers could bring sudden bursts of traffic, making demand difficult to predict. Jung and the engineering team used managed cloud services to handle that variability with limited development resources. He compared the components to Lego blocks: pieces that could be assembled into the system the business required.

It is an unusually domestic image for a large data operation. The pleasure sits in composing something useful from available parts. Building every part yourself would also be an activity, but the customer is waiting for the finished structure.

The Snowflake decision belongs to the same practical story. Jung’s interest included the cost model, the ability to try the service and easier access to data. These choices join product development to operating economics. A feature has to work; a company has to be able to afford delivering it repeatedly.

Fifteen minutes to find the trouble

Another episode makes the hidden work easier to picture. A sharp increase in tracking variations from an integrated advertising partner put pressure on a database and slowed the service. An alert led the team to the cause within fifteen minutes. Contacting the partner and completing the response took about an hour in total.

The sequel mattered as much as the diagnosis. The team added alerts for abnormal activity so it could respond earlier. Jung also described bringing operational information into the company’s messaging tools. Customer-facing colleagues could check useful data themselves, reducing the need to route every question through an engineer.

That changes the shape of a working day. A person answering a customer can investigate the delay rather than simply forward the complaint. An engineer has fewer interruptions. Information becomes useful to more of the people responsible for the result.

The episode offers a modest, concrete portrait of technical leadership. The work continues after the incident is closed. Someone has to improve the warning, shorten the next investigation and make the answer accessible. There is little ceremony in any of this. Customers generally prefer their infrastructure uneventful.

When the trail becomes harder to follow

Measurement has another complication: access to individual-level data can change. In 2023, Jung wrote about the renewed relevance of marketing mix modeling, a statistical approach that studies relationships between marketing investment and business outcomes. The technique predates the smartphone. Its renewed attraction came from a contemporary problem: how to make decisions when tracking provides a less complete picture.

His argument included a useful qualification. A statistical model has limits and will differ from reality. Marketers still need ways to judge spending and compare possible choices. A model offers a method for working with uncertainty; it does not make uncertainty disappear.

That year, Airbridge’s MMM Studio announcement identified Jung as chief product officer. The product combined model configuration, performance analysis and budget optimization. The progression is worth following: collect the information, interpret it, then help someone decide what to do with it.

“discover their true sources of growth”

Hunjae Jung, on the mission behind MMM Studio, 2023

The short phrase makes a larger promise than counting. A number can tell a marketer that something happened. Deciding how much credit belongs to an advertisement, and what to spend next, asks more of the system. Those questions help explain why the product role became such a natural extension of the engineering work.

Two people at the top

In April 2025, Jung became co-CEO. The division of responsibilities was explicit: Roi Nam would concentrate on sales and external activities; Jung would focus on products and internal operations. The arrangement arrived in AB180’s tenth anniversary year, after a long period of working together.

Nam’s welcome emphasized Jung’s curiosity and energy. He also described a mentor and a friend. The public account gives the appointment a personal texture: two colleagues with a shared history allocating the next set of jobs.

Shared leadership still needs a working design. Product decisions reach customers through sales and support. Customer requests travel back toward engineering. In this arrangement, each founder took responsibility for a different stretch of that traffic. Jung’s remit brought the company’s internal operation together with the products it was trying to build.

AB180’s current leadership description credits him with its core technology and notes his earlier CTO and CPO responsibilities. The accumulation matters. The CEO title sits on top of years in which the consequences of a decision would show up in a release, a bill or an alert.

The next system has to act

By 2025, Jung was publicly describing a move from marketing analysis toward AI agents. He discussed Luft, the company’s user analytics database, as part of the foundation. The ambition extended beyond displaying results to software that could participate in the work those results prompted.

At Modern Growth Stack in Seoul on July 21, 2026, his scheduled session was called “Stack AI, Rewrite MMP.” The title put the proposed change directly inside Airbridge’s own category: the mobile measurement partner. A company built to help people understand campaigns was considering how software could help execute them.

Jung’s presentation stressed data and organizational context. A useful agent needs to know the objective, priorities and constraints against which its decisions will be judged. Identical campaign figures can lead to different choices when the business goals differ. The numbers need instructions about what matters.

He also introduced Airbridge GO as a product in development, with a closed beta planned. Its proposed starting point was an app URL, followed by assistance across campaign strategy, creative, budgets and publication. That was the ambition presented in July, with development still part of the story.

Read alongside his earlier work, the proposal has a recognizable shape. First make a route into the app. Then make the journey measurable. Then help decide what should happen next. Each stage adds responsibility to the software, and another set of engineering questions to the person building it.

Jung’s career has moved with those questions. The job now includes a company as well as a product, and decisions as well as data. Behind the little click, the work keeps getting larger.