Who's this Klanger? An interview with Marijn Zwemmer, Unity Engineer
April 03, 2020
Marijn started developing games in the heart of Amsterdam almost seven years ago. After moving to Berlin and working at Wooga for three years he was up to a new challenge: Making the Seedlings come alive with AI!
He loves everything about game development, from game jams to low-level optimizations.
When did you start working in the gaming industry?
Phew, I actually have to look up the date on LinkedIn... I started as an intern at Little Chicken Game Company in Amsterdam in 2013.
What brought you to Klang?
SEED, of course! After leaving my previous company Wooga and travelling for a while, I was not sure what I wanted to do next. Klang caught my interest with SEED with its ambition and I joined quite quickly.
Tell us about your role at Klang.
I’m a game engineer on the AI team. Which means I’m having a nice balance between working on longer term AI systems and more gameplay related features.
What are you currently working on? Are there any specific goals, challenges, learning processes or results?
Right *now* I’m working on migrating all of our AI content to a new format, it’s not always flashy features… The past months but also still the coming time there’s a lot of work on combat. It’s been quite a challenge to fit all this in our current systems and not let our servers burn. Our game is quite slow paced overall, but then there’s all of a sudden a huge battle! From a technical/performance perspective that’s tricky.
What´s to be expected by Pioneers?
I expect (ok, hope for) a nice well-rounded experience where people will collaborate in all sorts of ways to keep their colony running and expanding. It’s going to be so interesting to see how groups will manage their colonies with their Seedlings!
What AI driven activities in SEED are you most fascinated about?
I guess it’s a part we haven’t done much feature work on yet but have plenty of ideas and inspiration for, which is relationships. It’s going to be so interesting to see Seedlings form relationships with each other for better or worse, and you as a player have to manage that somehow, also in communication with other players.
AI or not, what else do you enjoy exploring most in SEED, at the current state of the game?
I loved our combat playtests in which we had to get a simple economy going and fend of creatures. Was so intense!
What was Breakthrough Week? What did you focus on, what did you learn?
I created a “Ping wheel system”. They’re at least common in MOBA’s and shooters. You can place (temporary) markers on the world to communicate danger, resources or anything you like. When we did playtests with combat it was obvious how needed this was, it was a lot of yelling in the office trying to give directions (Creatures coming from the right! - Which right??).
In what way is SEED different from any other projects you worked on before?
It’s a lot more simulation-heavy than I’m used to. Previous games I worked on were easier to validate, for SEED this is a lot more tricky as a lot of the value comes from the emergent gameplay that will depend on the people playing SEED.
Did you ever give any talks yourself?
Publicly barely, I did a talk at Unite a couple years ago which is on YouTube. That was fun!
What´s your preferred source to deepen your knowledge?
Doing side-projects. I quickly get absorbed into something and will have to figure out everything, which means a lot of trying out and online research.
Could you tell us about one of your side-projects?
Hmm.. Over the past year I had a bunch of different things. I was prototyping a game in which you were a penguin having to throw other penguins off an iceberg. That led me into many months creating my own Navigation Mesh generation and pathfinding sub-project which I didn’t release yet. Some of the other things I did are on my GitHub, but there’s so many half-finished things..
What was your first project in Game AI?
I guess about 8 years ago, we made a game at university where we created a Navigation Mesh and Influence Map to do the pathfinding for AI enemies with Splatoon-like gameplay. It’s funny I work with these subjects on SEED a lot as well!
What could be a potential limitation or risk of Game AI implementation? Have you experienced any (as a Unity Engineer)?
Let’s see.. There’s this constant balance (which doesn’t apply to just Game AI) between working on more future proof/longer term systems and shorter term/validation of gameplay. It’s easy to go off balance in the first direction, creating systems that support cases that we’ll never have with unneeded complexity. But the other side is just as bad, having a patchwork of small hacky systems that are limited in what they can do and don’t work well together.
Can you describe SEED in five words?
Collaborative, Experimental, Endless, Explorative, Imaginative.
How do you see the world you experience in the game related to your own daily life?
I hope that SEED can become a place where people can try out different (political) structures and hierarchies. Most dreaming me would hope for people to have SEED as a way of iterating on this!