Develop and enhance generative models for:
- Text-to-motion: Converting textual descriptions into realistic 3D character animations.
- Style transfer: Adapting motion styles (e.g., neutral → zombie).
- Real-Time Motion Interpolation: creating movements for real-time applications
Implement and optimize modern approaches such as VQVAEs, diffusion models, auto-regressive models, and reinforcement learning for motion generation.
Develop pipelines to convert 2D videos into 3D animations using advanced techniques in computer vision, SMPL, and motion capture.
Build scalable pipelines for processing large-scale 3D motion data.
Collaborate with the founding team, animators, and game developers to tailor solutions and integrate seamlessly into workflows.
Translate research concepts into production-ready implementations.
We are looking for a highly skilled, collaborative engineer who thrives at the intersection of research and production.
Experience
- 6+ years in machine learning, with a strong focus on research, development, and generative models.
- MS or Ph.D. in AI, Computer Science, Mathematics, Physics, or another relevant field; or equivalent real-world experience.
Technical Expertise:
- Proficiency in PyTorch and familiarity with frameworks like TensorFlow or Keras.
- Hands-on experience in at least one of the following: VQ-VAE, diffusion models, auto-regressive models, or reinforcement learning-based motion systems.
Mathematical Foundation:
- Strong grasp of linear algebra and 3D geometry
Programming:
- Expert in Python
- Experience building scalable, production-level ML pipelines.
Gaming Enthusiasm:
- Passion for and experience playing video games, understanding player dynamics, and how animations enhance immersive experiences.
- Familiarity with 3D animation tools (e.g., Blender, Maya, Unreal Engine, Unity).
- Academic background with publications in top-tier conferences (e.g., SIGGRAPH, CVPR, NeurIPS).
- Experience working with 3D motion capture data, SMPL, or computer vision for 3D animation.
- Previous experience in early-stage startups or a fast-paced R&D environment.
