Microsoft Cambridge Research Intern – Generative Models Video Games 

United Kingdom, England, Cambridge 

871421697

This is an exceptional opportunity to work closely with a highly collaborative and interdisciplinary team. Some research challenges that we are currently tackling include, but are not limited to:

• Natural language control of embodied agents
• Test-time compute for multimodal agents
• Training and evaluation of multimodal generative AI models
• Generalization to new games/scenarios
• Model architectures/approaches for faster/improved world modelling
• Alternative approaches for video generation/understanding/compression
• Fine-tuning/adapting pre-trained models with minimal data/compute
• Deeper understanding/mechanistic interpretability of gaming-related foundation models
• Scaling to large-scale data and compute The focus and scope of each internship considers the team’s direction as well as successful candidates’ experience and research interests.

publications as an attachment.

Required/Minimum Qualifications:

• Currently enrolled in a Bachelors (BSc) or Masters (required) or PhD (preferred) program in Artificial Intelligence, Machine Learning, Computer Science, Game Development or a related area.

• Ability to carry out research in at least one of the areas mentioned above, demonstrated by at least one journal or conference publication (or demonstratable equivalent experience).

• Strong understanding of state-of-the-art deep learning approaches.

• Hands-on experience in implementing and empirically evaluating deep learning approaches.

• Effective communication skills and ability to work in a collaborative environment.

Preferred/Additional Qualifications:

• Currently enrolled in a PhD program in Artificial Intelligence, Machine Learning, Computer Science, Game Development or a related area.

• Ability to carry out research in at least one of the areas mentioned above, demonstrated by at least one journal or conference publication in one of the top publication venues in your research area.

• Demonstrated ability and/or strong motivation to learn to use cloud infrastructure for experimentation is a plus.

• In collaboration with your mentor and a diverse team (for example, including designers, engineers, and researchers), contribute to solving an ambitious research challenge and translate your results into actionable insights that are relevant to potential applications in modern video games.
• Write code and contribute to shared codebases to test the new approach or hypotheses.
• Distil the developed insights into effective communications, such as a research paper, prototypes, demos and/or a presentation, to reach internal and external technical and general audiences.