Zoe is definitely stands out in our cast. I think there are two reasons for this. One is her head and eye proportions, and the other is her extremely cartoony animations. For the head we settled on the final proportions in order to make Zoe look more like a child. When we tried normal size head proportions, many people were reading her as closer to Jinx/Lux in age rather than a kid. We ended up going with similar proportions to Annie as that made her read as a child much clearer.

As for her animations, we really wanted to push her to feel like the Trickster God archetype in game. To quote WIKI on the Trickster Archetype:

“The trickster crosses and often breaks both physical and societal rules. Tricksters…violate principles of social and natural order, playfully disrupting normal life and then re-establishing it on a new basis.

Often, the bending/breaking of rules takes the form of tricks or thievery. Tricksters can be cunning or foolish, or both. The trickster openly questions and mocks authority. They are usually male characters, and are fond of breaking rules, boasting, and playing tricks on both humans and gods.”

We wanted Zoe to on the surface appear human, but really behind the surface is a god motivated by its own self-amusement. We wanted her to feel like she was breaking and bending the rules of Runeterra—and even our own cast. Early on we even tried crazier things; at one point when Zoe used her R she came out of the portal on a roller coaster, which we ended up feeling pushed things a bit too far. In the end, though, we wanted her overall feeling, from VO, gameplay, and art, to push, bend, and break the boundaries of the game, like a true Trickster God should.

Many, but not all, interactions are known and assessed in advance for balance, interest, etc.

Sometimes, we build around them intentionally. The point of a summoner spell, item, or ally’s kit is to change something about how your champ plays. This often involves changing the nature of a spell. Indeed, these changes are almost inherently more interesting than simply amplifying a spell (as AP tends to do).

Sometimes, we spot them emergently in internal testing. Most Rioters are experienced League players—we’ve been playing for years. We tend to understand the probable ways to break a kit that have been seen before. Even when unplanned, if spotted, the team can backtrack and re-design the interacting spells to be balanced/fun/compelling/etc.

Sometimes, you all surprise us. This is particularly true of champion interactions, since there are so many (there are currently 45 billion unique five-champ compositions available). In this case, we often have to go back and evaluate how we want the interaction to work, but we approach the situation with a strong bias toward supporting the strategies players have discovered.

Once identified, we ask whether we believe we can support the interaction. If we can’t, we need to either disable or modify it. This is a cost/benefit analysis: Is the cost of combining or investing in the two elements of the interaction sufficient to make the combo balanced?

For example, the Bodyslam+Flash combo yields a dramatic increase in reliability. It’s honestly a degenerate interaction. However, it involves Flash, so the massive cooldown/power level makes the opportunity cost of using the combo very high to compensate for its strength.

Fiddle’s Crowstorm+Zhonya’s combo gives a dramatic reduction in risk, but actually reduces the reliability of Crowstorm, since by Hourglass-ing, Fiddle roots himself. Typically, these are the most interesting and long-term healthy interactions—where the costs and benefits are actually internal to the spell being affected.

A recent example of this actually comes from Zoe. We enjoy supporting Flash + cast time spells (e.g., Alistar Pulverize+Flash/Ahri Charm+Flash) to increase reliability. We thought it would be fine to support it on Zoe’s E, Sleepy Trouble Bubble. However, as she has access to more than six times the Flashes of a normal champ, the typical rationale for that support (the long CD and opportunity cost of burning Flash to increase the reliability of an offensive action) became inapposite. Unfortunately, I came to this conclusion very late in her development cycle, which left few elegant solutions. In the end, I decided to lock out Zoe’s Flash during E’s cast in order to allow her to be about more than setting up E+Flash plays. After all, why would you cast a risky E over a wall when you could just guarantee the hit?

Perhaps we will revisit that decision in the future, either by providing a less artificial blocker to the interaction or by re-assessing whether Zoe’s ability to frequently set up E’s with Flash is actually unhealthy. Such is the benefit of making a continuously updated game.

Riot’s always looking for engineers to level up our tech, and we’ve got a couple top priorities when it comes to finding the right fit. First, we’re interested in T-shaped engineers. That means we value breadth and depth of skill. The technical landscape at Riot changes quickly and we love engineers who are willing to adapt, learn new tech stacks, work with new teams, and break out of their comfort zone to deliver products to players. Second, we value engineers (and Rioters in general) who thrive in a feedback-oriented environment. In order to maintain our culture, deliver the best possible tech, and keep our focus squarely on players at all times, we need engineers who can give and receive tough feedback.

Beyond that, requirements vary from role to role, team to team, and even day to day. You can see our rubric for evaluating growth in the engineering discipline in Mike Seavers’ Tech Blog series on debugging titles, and our engineering careers page contains more insight into how we think about interviewing for engineering roles.

At the end of the day, we hire passionate gamers who put players first. If that sounds exciting to you, you’re on the right track.