Who's this Klanger? An Interview with Alessia Nigretti
February 17, 2020
Graduated from a Computer Science and Artificial Intelligence degree and with experience as a Technical Evangelist at Unity, Alessia contributes to the AI systems of Seed. Coming from a multidisciplinary technical background, Alessia has developed biofeedback-enhanced audio-visual tools for virtual reality mindfulness and AI-driven digital art installations displayed across the globe.
Why did you start working in the gaming industry?
I started working in the games industry because I have been interested in games since I was a little kid. I remember watching my mum play Tomb Raider on her PS1, and I remember being really attracted by the visuals and the challenges. Growing up, I started questioning things like “How is this made? How can I make it by myself?”. At school I was good at maths, and was balancing homework and a lot of gaming, so I thought that game development could be my direction.
Where did you study then first?
I studied in Brighton, UK, and my course was a Bachelor of Science in Computer Science and Artificial Intelligence at the University of Sussex.
What brought you to Klang?
I was still in my last year of university when I came across Klang and was fascinated by their ambition. I studied AI with the goal of getting into Game AI, and I realized SEED was aiming really high on that front. I started researching and finding out more about it and decided to get involved!
Tell us about your role at Klang and a bit about the team you are in.
My role at Klang is Game Engineer and I work in the AI Team, or “Team Activities”. We are in charge of the decision-making process behind the AI agents in SEED, which is handled by our Utility AI system. Practically, we make sure agents pick and execute activities in a smart and efficient way.
What are you currently working on at Klang? Are there any specific goals, challenges, learning processes or results?
I am currently working on optimization, taking care of the execution part of all the activities that are in the game. In particular, because of the load on our server that is expected for Pioneers, I have been working on scalability and performance. It’s been quite a challenge so far as I had never worked with systems of this scale, but it’s been really interesting to figure it all out.
What is the number you are expecting for Pioneers?
We are planning to be able to support roughly 300 characters, 150 creatures and 50 buildings in the game with stable performance, which is actually more than we are expecting for Pioneers.
Could you name some examples of activities you are talking about?
Everything that the agents in the game can do is an activity. Whether they are wandering around or they are consuming food or sleeping, they are executing an activity.
Activities are handled in SEED using Utility AI - a system that, in our implementation, is based on three phases: discovery, selection and execution.
Agents look at everything that is available for them to do in the world, then filter out what is not possible for them to do, select what is best for them depending on their current state, and then execute the selected activity.
Were you always interested in Game AI?
I´ve always been interested in anything AI related, and because I was passionate about games I really wanted to see how well these two things could combine. I think game AI is a great way to make a game more immersive. When I joined Klang, I didn’t know about Utility AI, but I dived into it thanks to my team and to Dave Mark’s talks.
Could you explain a bit more about SEED as a data-oriented game?
In SEED, we constantly deal with a large amount of data. For instance, for each character we have different attributes that change the way they interact with the environment, and these are all being stored as data. This data can however change during runtime, so we need to work in the most optimal way to make these changes as fast and efficient as possible. We are optimizing for scalability, which means that when we add content to the game, we expect the system to be strong enough to be able to handle it.
What was Breakthrough Week? What did you focus on, what did you learn?
Break-through is a week used to explore features or concepts that are experimental or not yet prioritised by design. During the latest Breakthrough Week I took inspiration from The Sims and I tried to create a system that would allow Seedlings to interact with each other. This was a feature that I was really interested in but hadn’t been prioritised for Pioneers. I started by narrowing it down to what I thought could have fit within the scope of the week and the resources I had. I really wanted our Seedlings to show some personality - I wanted to see what the game would look like if we tried to make our characters feel somewhat more human, for example seeing Seedling mourn the loss of a loved one.
During this week I was not working with my regular team, but instead with Paula (animator), Hossein (artist) and David O (game designer). After Breakthrough Week we showed our work to the rest of the team and managed to get it prioritized and included as part of Pioneers.
It was interesting to work with people I do not normally get to work with. During the week, nobody really had a fixed role, and we were all contributing to the same idea using our skills in the best way we could.
Could you tell us a bit about the Code Talks happening at Klang? And an example of inspiration you gained through this occasion?
We have Code Talks every Thursday morning and usually one Engineer in the group has a topic that they want to share with the rest of the group. One of the topics that has been the most useful for me was a talk given by one of our Game Engineers, Ulrich, a few weeks ago, about scenario tests - a type of test that is useful to measure performance of a specific scenario or feature. This was really useful for me to construct the scenario for our OKR goal - the 300 seedlings, 150 creatures, 50 buildings. This type of testing makes our measuring a lot easier.
We just recently started to change the Code Talks a bit, and last week we had our first Lightning Talks: Here we have short 10-minute talks given by different engineers. The talks may also not be strictly related to SEED - the themes can be very different. Talks can be about any ongoing or past side-projects, or just a knowledge-share talk for the rest of the team.
Have you recently worked on any side-projects?
One of my most recent side-projects was a virtual simulation of genetic evolution at a microbiological level. I was working with Sascha Pohflepp, art director, and Matthew Lutz, computational biologist, on a virtual installation that got exhibited in the Garage Museum of Contemporary Art in Moscow, check it out here. We worked on that for about two months, and it was very cool to put together knowledge from very different fields to build something so unique.
What´s your preferred source to deepen your knowledge?
I like to attend conferences, mainly game-related, because I really like the environment - I think there´s always a lot to learn. People are usually working with games but can come from many different backgrounds. One that I really like is Gamescom in Cologne, happening yearly in August - definitely worth a visit!
You've been to many conferences to hold talks about your skills, what would be your next talk on?
I delivered some talks in the past few years. I was a Technical Evangelist at Unity, exploring a lot of the features that were still work-in-progress, and interfacing with the public. Thanks to that role, I realized that I really enjoy talking to people and sharing knowledge at the best of my skills! See one of Alessia´s past talks as a Unity Evangelist here.
I have plans for the end of March to teach a class at the CIEE Global Institute in London about my experience going from university to the games industry, but I am also keen on talking about all the cool stuff we are doing with AI, so I am keeping my eyes open for conferences and events!
What are you thinking about when you see the Seedlings going on their daily business? On this new planet?
It´s a bit difficult to answer this questions because from the moment I started making games myself it’s been very different to also play them, as I am always thinking about everything that is happening on the code-side. It´s kinda difficult to interface with what I am seeing and just enjoy the experience. Whenever I join a SEED playtest I am always thinking about finding what’s going wrong. So I find it hard at the moment to completely abstract that away and think about the purely gaming aspect.
Is that something you would say many Klangers experience around Playtests these days?
I think many of the Engineers do, yes. I am hoping from the moment we have interactions in game, this will become less of an “issue” for us in a way.
Do you remember some of the first impressions you had of SEED?
Before joining Klang, I was really impressed by the ambitions of the project, and all the attention that was put on the “emotional” side of it. I also really liked the low-poly style and the aesthetics in general.
When I joined and I tried a SEED playtest for the first time, however, I was admittedly quite confused! At the time we weren’t paying much attention to User Experience and User Interface, and it’s cool to see how much has changed now.
Is there anything you know that is going to be on your plate in the future, maybe not for Pioneers, maybe later - something you already see as a challenge or potential excitement coming up?
There´s a goal that has been on my plate since I joined Klang and it’s AI planning, which will probably be our own implementation of Goal-Oriented Action Planning (GOAP).This has been a challenge for SEED for quite some time because of the complexity of the systems we have. I have been doing some prototypes of this, and even if it’s probably not in scope for Pioneers, it´s something I am really passionate about. Some of the work I have done on it is coming straight from my studies, so I want to explore it a bit further and see how it falls onto an actual project.
What are you thinking about Avesta?
I really like the settings, the whole lore of the game as we are setting it up. I really like the parallelism with the real world, how we introduce SEED to someone who doesn't know about it. Whenever I talk about it, I ask to imagine that Planet Earth has been destroyed, because of everything humans are doing to it (think about global warming), and that a few survivors have been sent to space in the “Seed Pods”, as we call them. They land in this new world, Avesta, and they start settling it from scratch, and using the technology they managed to export. I think Avesta has a lot of space for some interesting explorations, in the sense that because of sandbox-y vibes of SEED, we are giving players the freedom to build whatever they want, kind of write their own story.. I am very curious about seeing what comes out of it.
You can check out the SEED Pioneers trailer!
Do you believe in “the” end of the world? What does it mean to you, as a future-scenario thought, a “consequence”, a concept, a myth?
I don't know how far it is. But I do believe that the end of the world will come because it does feel like we are caring less and less about the ecological conditions of our planet. I don't think it´s a myth or a concept, but a fact.
Regarding the interesting parallelism you mention, between the world you experience with SEED and your daily life - Will there be a sort of “fact checking”, or knowledge systems implemented in the game to be reminded of the reality outside the game?
We are making a simulation, and we are trying to make it as realistic as possible, taking into account that we are on a different planet and in a far future. We want players to use their real knowledge of the world, to create something that is as similar as possible to the planet they live on - but hopefully less messed up!
I am looking forward to seeing more collaboration-based systems in SEED. I want to see how much of a community-driven game we can make it. I am interested in seeing how many players will be collaborating towards the same goal or how much someone is going to be playing the part of the “bad guy” and go against everyone else.
But it's possible someone does that?
I think so! I mean if a player or a character doesn't work as a team, as a community, they are going to be penalizing the rest of the group. Whether that will take them somewhere - that is up for them to figure out!
Is there something you would love to be possible in the world of SEED sometime soon, any specific experience you dream of as a player and then as game engineer?
I am really interested in seeing how many players are going to be able to figure out whether they are playing with a player that is behind the screen or with an AI that we have programmed. I don´t know if there´s going to be a way to tell players whether they are interfacing with a playing character or not, but it would be cool to make the difference between the two as thin as possible.
What about the theme of combat in SEED?
We have the concept of combat already, and we have the concept of defense, too. We have creatures - they are still a prototype at the moment - and you can place turrets around your colony to defend your Seedlings from them.
Deepening the thought of a parallel world: The topics of nationalism and borders for example are of course very controversial and complex in this sense.
There are many decisions facing these themes that have not been made yet. For Pioneers we are going to have a community working together and then creatures that are attacking, so we don't really have factions yet. But that is definitely in the planning. And once that's in, we are going to start seeing the effect of people choosing whether they want to work together or not..
The creatures in Pioneers are AI driven?
Yes. There are both hostile and friendly creatures. But we are exploring different behaviors. So this is still highly work in progress. But there´s going to be some variety for sure.
Do the creatures have their own settlements, e.g. if they do not find a colony to attack do they live for themselves remotely?
Their own wilderness? If a player chooses to venture out a bit too far outside of the colony, yes, there is always the risk that you are going to encounter these creatures. If the Seedlings go on their own, for example if they decide to not collaborate with the rest of their colony, they might risk their life. Venturing out too far, possibly not strong enough for what they will find..
This confrontation with creatures, would you say is one of the main challenges in Pioneers?
Yes, it´s going to be one of the main challenges. You are going to have to think smartly and work together to use your resources in a way you can defend your colony from what will come from the outside.
Creatures you collaborate with in your daily life in your colony?
We have friendly creatures. We just call them “Dogs” internally for now, because they are always good - they walk around Seedlings, follow them in their daily life and follow each other around. You can differentiate them from the ones that might attack you. And it´s only one of the first versions of creatures that we are planning to have, there´s gonna be more!
Can you tell us something about the subject of sustainability in SEED? E.g. Environment protection and green technologies?
We have one system in the game that is based on chopping down trees to obtain resources from them. Trees like in real life are finite resources, so you are going to have to be careful in how you use these resources you are given in the game, not wasting them, like in the real world! So I guess there´s attention to sustainability, and there will be more exploration on that very soon.