Game systems
Designing loops, progression, and social features that remain legible at live-game scale.
About
I connect design intuition, quantitative evidence, and engineering craft to build systems people can understand and feel.

Levon Zhao
Game designer → software engineer
The through-line
I spent more than three years designing live game systems at IM30 / Tap4fun before moving deeper into computer science, machine learning, and software engineering. The medium changed; the core question did not: how does a system communicate its state, reward understanding, and improve through feedback?
My game-design work supported a strategy title generating more than $20 million in monthly revenue. One featured design increased player engagement by 67% and reached a 157% higher participation rate than concurrent features. SQL and Python analysis helped the team monitor behavior across tens of thousands of players and turn the findings into repeatable reporting practices.
Graduate work at Northeastern and Georgia Tech broadened that foundation across algorithms, systems, AI, graphics, and game development. I then built production AI and analytics systems at Activision Blizzard and now evaluate frontier models through Handshake AI. This portfolio keeps both sides visible: the product judgment behind the work and the implementation details that make it real.
Across disciplines
Designing loops, progression, and social features that remain legible at live-game scale.
Using SQL, Python, playtests, and experiments to turn player behavior into the next design decision.
Building machine-learning, graphics, and product systems around a clear human interaction.
Beyond the résumé
I'm a certified professional Mahjong player in China, have passed the CFA Level I exam, and own an automatic Mahjong table that has become a surprisingly effective systems-design lab. Probability, incomplete information, and decision quality show up in games, finance, and engineering in different clothes.
A few frames

