Neural Model Game Engine Developer (GameNGen)

Save to Favorites

We will be competing at the NVIDIA developer contest: https://developer.nvidia.com/llamaindex-developer-contest

The code to implement the GameNGen is here - https://docs.google.com/document/d/1Mr1FQt-UO9KqNd9gpod3sju2L3mxyaoLEJItsnXYj-E/edit?usp=sharing

About Us:

Our organization is at the forefront of cutting-edge AI research and development. We are developing GameNGen, the first game engine powered entirely by a neural model. GameNGen enables real-time interaction with complex environments over long trajectories at high quality. This groundbreaking project simulates classic games like DOOM at over 20 frames per second on a single TPU, with next-frame prediction quality comparable to lossy JPEG compression.

We are looking for passionate volunteers to join our innovative team and help bring GameNGen to life.

Position Overview:

As a Neural Model Game Engine Developer volunteer, you will be integral in developing, testing, and optimizing GameNGen. Your primary focus will be on training reinforcement learning (RL) agents, fine-tuning generative diffusion models, and implementing real-time frame generation. This is an excellent opportunity to work with advanced AI models, including Stable Diffusion, and contribute to pioneering technology in the gaming industry.

Key Responsibilities:

- Reinforcement Learning Agent Development:

- Train and optimize RL agents to play classic games, generating high-quality training data.

- Implement and refine neural network architectures to improve agent performance.

- Generative Diffusion Model Training:

- Fine-tune diffusion models for next-frame prediction, ensuring visual stability and quality.

- Implement conditioning augmentations to stabilize long-term autoregressive generation.

- Latent Decoder Fine-Tuning:

- Work on the latent decoder of Stable Diffusion to correct artifacts and improve image quality.

- Performance Evaluation:

- Evaluate model performance using metrics such as PSNR and human evaluation.

- Contribute to the generation of full gameplay videos showcasing GameNGen’s capabilities.

- Collaboration and Documentation:

- Work closely with other volunteers and researchers to ensure smooth integration of components.

- Document your work to support ongoing development and future scaling of the project.

Qualifications:

- Technical Skills:

- Experience with deep learning frameworks (e.g., PyTorch, TensorFlow).

- Knowledge of reinforcement learning and neural network architectures.

- Familiarity with generative models, especially diffusion models like Stable Diffusion.

- Proficiency in Python programming.

- Experience:

- Prior experience in game development, AI research, or related fields is a plus.

- Understanding of image processing and evaluation metrics (e.g., PSNR).

- Soft Skills:

- Strong problem-solving abilities and creative thinking.

- Excellent communication skills and ability to work in a team.

- Self-motivated and able to work independently with minimal supervision.

What You’ll Gain:

- Hands-on experience with state-of-the-art AI technologies.

- The opportunity to contribute to a pioneering project in the gaming and AI space.

- Collaboration with a team of passionate and innovative AI researchers and developers.

- The potential for future involvement in larger projects or a transition to a paid role.

How to Apply:

If you’re excited about the potential of neural model-powered game engines and have the skills to contribute, we’d love to hear from you! Please send your resume and a brief statement of interest.