Three students entered a competition they knew little about. Today, they’re helping shape how millions of people see ads online.
Zhang Chang, Li Qiang, and Wang Zhaobing are colleagues in the Tencent Marketing Solution (TMS) department, but their journeys began several years ago in an unexpected way: the Tencent Advertising Algorithm Competition. Before entering the competition, none of them had experience in advertising or saw it as a potential career path.
While they were among the first to enter, the annual competition attracts thousands of students from around the world. More than just a test of technical skills, it’s a gateway to applying classroom knowledge to real business challenges with impact. Students work in teams over several weeks, using anonymized data from Tencent’s ad systems to solve problems that affect millions of users daily.
In this article, Zhang, Li, and Wang reflect on how stepping outside their comfort zones led to meaningful careers, and why this competition became a turning point in their professional lives.
Zhang Chang: Predicting what people want, when they want it
2017 | Top20 | Now: Algorithm Engineer, Tencent Marketing Solution
Zhang Chang was studying computer science when he joined the 2018 competition, which he saw as a chance to test himself on something new.
The challenge: how to predict whether someone will click on an ad when their interests constantly shift throughout the day.
Zhang and his team designed models that could understand how people’s tastes change during the day. For example, someone might search for coffee in the morning and dinner options in the evening. They also combined multiple prediction models, like gathering opinions from a group of experts, to make their predictions more reliable. This smart mix of data cleaning and thoughtful model design helped them place in the top 20.
“Before the competition, I had no background in advertising. But through it, I began to understand the core goals of the industry, to make accurate, efficient decisions at scale.”
After graduating, Zhang joined Tencent in 2019. Today, he builds smart systems that help show the right ads to the right people at the right time. One idea from the competition – using “freshness” to measure how new or relevant a product is – was later added to Tencent’s actual ad system and boosted click-through rates by two percent, an innovation in the company’s advertising business.
His advice for future participants? “Don’t worry if you’re unfamiliar with advertising. Dive into the data, think creatively, and bounce ideas off your teammates.”
Li Qiang: Solving the timing puzzle
2017 | Second Prize | Now: Algorithm Engineer, Tencent Marketing Solution
While Zhang focused on shifting interests, Li Qiang tackled a different puzzle. As a master’s student at Dalian University of Technology, he joined the 2017 competition with two teammates, forming a team called “Raymone”, a nod to basketball star LeBron James.
Their challenge: people don’t always make a purchase right after clicking an ad. So how do you build a system that can predict purchases when people might not buy anything until days later?
To solve this, the team grouped apps based on how long users usually took to make purchases. For apps with longer delays, they removed recent data that likely didn’t include completed purchases. For faster apps, they only trimmed a few hours. These timing adjustments made their model more accurate, earning them second place.
“We didn’t know anything about advertising back then… But analyzing data revealed patterns that helped us bridge the gap.”
Li joined Tencent soon after the competition and built a system that helps advertisers get better prices when placing ads – this system now handles tens of millions of requests every day. Today, he leads work on models that look at how people’s behavior changes over time.
For those worried about a lack of experience, Li says: “If you’re willing to learn and think carefully, you can do it.”
Wang Zhaobing: Connecting the dots over time
2017 | Top 10 | Now: Platform & Content Ads Algorithm Team Lead, Tencent Marketing Solution
Wang’s approach differed from both his future colleagues. As a master’s student at the University of Hong Kong, he joined the inaugural competition solo in 2017, mainly hoping to learn something new.
Instead of looking at each user click or conversion as a separate event, Wang mapped out how people’s interests changed over time. By tracking these patterns, he was able to make better predictions about future behavior. He also built a tool to automatically test and adjust his models during the competition, earning a top 10 finish.
This method, called user behavior sequencing, was later used in Tencent’s advertising systems. Today, Wang leads a team that helps improve how ads are delivered across Tencent’s platforms like Video, News, and streaming services.
“This competition was like a bridge connecting academic knowledge to real-world applications.”
Wang credits the training and mentorship systems at Tencent for his promotion to managerial positions. “My mentor helped map out a growth plan and I’ve followed that path since.”
Opening doors, changing futures
For Zhang, Li, and Wang, the competition revealed career paths they never imagined. Their stories show that competitions like this offer more than prizes, they offer perspective and open doors to unexpected opportunities.
“I didn’t plan for this path,” Li Qiang says. “But looking back, I wouldn’t change it.”
Want to join?
Registration for the 2025 Tencent Advertising Algorithm Competition is now open. This year’s challenge focuses on frontier AI technology: multimodal generative large language models. Students receive mentorship from Tencent engineers and compete for prizes, internships, and potential job opportunities.
One of this year’s participants, Nova, is excited about the competition, saying: “The challenge in this competition is all-modal, and completing just one scenario will be a huge advantage for fresh graduates looking for a job. That’s the real significance of it.”
Deadline: July 31, 2025
Register now: https://algo.qq.com