AI coding models are changing rapidly, but choosing the right model for game development isn't as simple as picking the model with the best benchmark score.
Different models have different capabilities when it comes to understanding game logic, writing code, debugging, working with large codebases, and completing complex tasks. A model that performs well on a single task may struggle when asked to repeat that task dozens or hundreds of times.
Vladimir Glazachev, CEO of Rosebud AI and an ML engineer with years of experience working with AI systems, will explore how to evaluate coding models specifically for game development.
- What should you actually test when evaluating an AI model?
- How do you compare models for different game development tasks?
- How do you push models beyond their default capabilities?
- How do you test whether a model continues performing over repeated tasks?
- What should developers expect from AI coding models as they continue to evolve?
Learn how to evaluate AI models based on what actually matters when building games, rather than simply relying on benchmarks or hype.
Session speakers
Learn more about the person leading this session.
Vladimir Glazachev
SpeakerCEO · RoseBud AI
Choose one admission option
Choose any available ticket or pass below. Free and paid options are both shown when offered.
Purchase-window dates are shown in your local time.