TikTik

I led a team of three on this project. We divided scope clearly: one colleague handled visual effects and UI; another built the particle system and a replay system that gave researchers a visual timeline of each session. I handled core gameplay systems.

Local multiplayer in Unity requires careful input routing: each controller needs its own player instance with no shared state. I built the input layer from scratch, handling controller assignment, disconnection recovery, and split-screen coordination. Level data is serialised to JSON and loaded at runtime, which meant building a level editor that non-programmers on the research team could use to create new puzzle configurations without touching code.

The replay system was the subtlest challenge. It needed to be accurate enough for a researcher to reconstruct exactly what each player did and when: frame-level event logging, not a video. Getting the data model right so it was both faithful and compact took several iterations.

The project demonstrated what it means to build for a dual audience: players who need to be engaged, and researchers who need the data to be clean.

Built for a cooperative decision-making study at the University of Geneva, where pairs of participants play while their choices are recorded for analysis.