A few pieces I can show, with the thinking behind them: features I owned, heroes I designed, and the engineering I started in. I'm currently a Senior Systems Designer on Horizon Hunters Gathering at Guerrilla (in development).
One player gets a live editor for the match while everyone else fights inside what they build. I owned the mode. The hard part is authority: enough power to shape the fight, not so much that everyone else is a spectator. It pushed user-generated content forward in Arma, a franchise that pioneered it.
A HUD has to work on a couch two meters from a TV and on an ultrawide up close. So I didn't build a layout, I built a framework: elements scale and give way by what matters most in the moment, with no fixed positions. I wrote it from scratch in Enfusion's proprietary scripting language, and I built the WYSIWYG editor on top so designers could tune HUD behavior live instead of waiting on a build.
The hard part wasn't drawing a HUD. It was the contract. A framework is only worth as much as what other people can build on it without breaking it, and here "other people" meant two groups I didn't control: the designers tuning it live, and the modding community extending it after launch. That is an interface problem, not a UI problem. Get the seams right, hide the internals, and keep the surface stable, and people build things you never planned for. Get them wrong, and every change you make downstream breaks someone.
Two things tell me the contract held. The code ships with the game, so the modding community reads it and builds its own UI directly on the framework, by the thousands, without it falling over. And it has been live for over two years with no architectural rewrite: the interface stayed stable enough that the insides never had to be torn out.
You don't have to take my word for it. The framework ships in the Reforger code, so you can read the real thing.
The longer version of how I think about interfaces like this is in Contracts Between Game Systems.
Four heroes I designed in a small pod, all built on one shared ability framework, so every kit had to play by the same rules as the rest of the roster while giving the team a different job.
Closing distance in an extraction shooter is usually a mistake. Taurus makes it pay off. Bull Charge punches him through the enemy line; Power Pylon, his deployable shield, turns wherever he stands into ground the team can hold. A brawler whose mistakes are his own: positional and readable.
Support falls apart when healing is a separate job from fighting. So Trenchwalker puts the sustain on the gun: the Leech Gun pays health for dealing damage, making aggression and survival the same move. The Hybrid Grenade heals allies and hurts enemies on one cooldown. Lifeline and threat, no tool-swapping.
Looks like an ambusher, plays as space control. Blink gives a short invulnerable reposition, so you beat him by predicting where he has to go, not out-shooting him. Plasma Arc punishes players for bunching up. Every exit planned before the entrance.
Weaver holds ground. His Energy Barrier turns positioning into a team resource, deciding which sightlines and fights exist at all. The Minigun rewards committing to that ground with heavy damage, at the cost of mobility. The skill is where and when you drop the barrier. The hero a team builds around.
My job here was engineering: multiplayer, the launcher backend, and the hero and ability frameworks designers built on. It's where I learned what designers do to frameworks built without them in the room.
A narrative adventure where you play the family cat, point and click. I handled the gameplay side: the NPC AI, the cat's point-and-click movement, mission logic, a stack of gameplay mechanics, and the UI code.