This job might no longer be available.

Machine Learning Engineer Internships

Craft:

Job Description:

Machine Learning Engineer Internships

There are 3 Machine Learning Engineer internship opportunities defined in our AI R&D and New Games teams. The goal of these roles is to (1) build tools and frameworks that makes machine learning accessible and understandable to be used in different teams and use cases at King, and (2) to scale modelling and make it part of game development automations.

Here, at King, we care deeply about the quality of game content and player experience. This is reflected in the AI research that we have conducted in the past few years. Previous AI powered content production automation work conducted at King include: Human- like playtesting deep learning models , capturing local and global patterns in procedural content generation via machine learning in collaboration with modl.ai, and generalized reinforcement learning for gameplay .

Please indicate in our application which project you are more interested in.

Skills to Create Thrills:

• You are writing your master thesis or have recently finished a master degree in Computer Science, Mathematics, Engineering Physics, Electrical Engineering, Industrial Engineering or similar fields. PhD students with related background are welcomed to apply.
• You have a passion for the topic.
• You are interested in reading research methodologies and implementing them.
• You have taken courses in machine learning and deep learning.
• You can demonstrate strong communication skills.
• You have strong programming skills in python (Experience with tensorflow or keras is a plus).
• Experience with tensorflow or keras.
• You care about code quality and are familiar with software development best practices.
• You have taken software engineering related courses and are familiar with microservices.

Required materials to apply:
Please put your (1) cover letter, (2) CV, and (3) transcripts in one single PDF document and attach it to your application.

Due to COVID uncertainties, we are currently considering only candidates that are based in one the countries where King has an office including Germany, Sweden, Spain, and UK.

1. Interpretability and Explainability of AI in industrial applications - AI R&D

One of the objectives of the AI R&D team is to enable model development and adoption of ML models in our products. At King, we care deeply about the quality of game content and player experience. Hence, it is important for us to understand model behaviors, track unexpected behaviors and the effect of models. A lot of advancement in AI has happened in the past few years with the help of more complex methods. There are several works published in the topic of explainable AI and interpretability of models (eg. Why should I trust the prediction of any classifier?). Moreover, a few frameworks have been developed (e.g. SHAP and LIME ) This internship work includes a study of existing approaches in the domain of interpretable AI, and developing a framework integrated into our modeling work at King. We look at methods to explain what neural networks learn, ablation studies, and investigate the suitability of existing explainability tools for our use cases in game development.

The work in this internship involves:

• Research on ML model explainability and interpretability.
• Understanding machine learning use cases at King.
• Developing an explainability component to train ML models as part of our ML platform.
• Developing an interpretability component as part of our ML product tools.
• Collaboration with engineers and researchers.

Required Skills:

• You are interested in data exploration and visualization.

2. Game content understanding and clustering - AI R&D

At King, we care deeply about the quality of game content and player experience. Hence, one of the objectives of AI R&D at King is to build AI- powered content production pipelines that enable automatic playtesting and improve the process of content generation. We have developed a content automation platform that enables research in this domain. As part of this initiative, the team has built a supervised learning model in its production pipeline . In addition, to create AI solutions to generate content, a good understanding of content and metrics that defines quality of content in relation to player experience is a key component. This internship focuses on understanding content and is defined in the two folds: (1) research to identify features and embeddings that describe the content the best to build models down the line, and to apply deep clustering methods to segment content. (2) development of data exploration and interactive visualization tools to integrate this solution to the content automation platform.

The work in this internship involves:

• Research and content generation and game playtesting.
• Research and development of data embedding and clustering methods.
• Implement clustering on a game content use case.
• Implement an interactive visualization to explore data in the embedded space.
• Integration of the developed solution into the content automation platform.
• Collaboration with engineers and researchers.

Project specific required skills:

• You are interested in data exploration and visualization.

3. Content Generation Framework - New Games

In the New Games studio, automation is at the heart of what we do. At King, we’ve been investigating techniques to leverage AI for content production automation. There are two steps when it comes to producing content generation and validation. We have been working on both parts of the pipeline, and our journey so far is nicely summarized in this talk presented at the Foundations of Digital Games Conference 2020. To assist content producers, the next step is to bring the state of the art procedural content generation to production by building platforms and tools that enable content exploration, model development and evaluation of generated levels using deep learning techniques. This internship will be centered around building such a platform for content generation, in this case levels, for games in development. There are a few approaches that can be looked at for the core content producing part of the framework such as generative adversarial networks (GANs), variational autoencoders (VAE), genetic algorithms (GE), etc.

The work in this internship involves:

• Research on procedural content generation state- of- the- art techniques.
• Building a framework for content generation and evaluation.
• Contribution to King’s research and production tools.
• Collaboration with level designers, engineers, and researchers.

Project specific required skills:

• Experience with Unity is a plus.

Want to join our Kingdom? Apply now! All applications need to be in English. This will be a 6- month full- time internship starting around September and is suitable for those who are writing a master thesis or recently finished a master degree. PhD students with related backgrounds are welcomed to apply. Our application deadline is April 18th 2021, however we recruit on an ongoing basis, so early application is advised to avoid disappointment.

Please note that this internship will likely be carried out on a remote basis. We are keeping this under review, and will provide final confirmation closer to the date on which the internship is due to commence, but you should assume the internship will be remote. To be considered for this opportunity you would need to be a resident of, and eligible to work in, a country where King is located, as King can only facilitate local hires in order to comply with its legal and tax obligations. These countries are: UK, Spain, Sweden, Germany, Malta. If you are resident and eligible to work in one of these countries you are free to apply for a remote internship in any of the other King locations listed here (even if it’s not the country you’re resident in). Please note that we will require evidence of your residence and eligibility to work as part of the internship application process.

About King