We are building the next generation of generative AI for 3D animation. At Animatica, we move fast to bridge the gap between cutting-edge AI research and practical tools for creators. We need a Machine Learning Researcher who wants to get their hands dirty advancing state-of-the-art generative models for human motion. You won't just be writing papers; you will be working directly with massive mo-cap datasets and collaborating closely with our ML engineers, animators, and graphics experts to build tech that actually ships.

What you'll do

• Designing and running experiments using models like Transformers, Diffusion, VAEs, and GANs to crack complex tasks like motion prediction, interpolation, style transfer, and trajectory generation.
• Wrangling large-scale motion capture and skeleton-based datasets (like AMASS, Human3.6M, Mixamo) and tackling real-world motion representation challenges using Euler angles, quaternions, and FK/IK.
• Digging into the latest research from top-tier venues (CVPR, SIGGRAPH, NeurIPS, ICLR, ECCV), implementing the best ideas, and adapting them to directly upgrade our core technology.
• Teaming up with our ML engineers to navigate production constraints, and working with the animation and graphics crew to evaluate data and integrate your models into the pipeline.
• Formulating hypotheses, documenting your breakthroughs, and contributing to scientific publications to put Animatica on the map in the global research community.

What we're looking for

• An advanced degree (PhD or Master's) in AI, Computer Graphics, Robotics, Bioinformatics, or a similar field, backed by at least 2 years of real-world research experience in academia or industry.
• A rock-solid theoretical understanding of machine learning, deep learning, and time-series modeling.
• High proficiency in writing code for modern ML libraries, with a heavy preference for PyTorch (though we will definitely talk if you are a TensorFlow or JAX wizard).
• Deep knowledge of how motion actually works under the hood, including local vs. global coordinates and forward/inverse kinematics.
• The ability to design rigorous experiments, analyze the results without bias, and clearly communicate your findings through technical docs and papers.

Nice to have

• Published papers on motion in-betweening, 3D pose estimation, procedural animation, or deep generative models.
• Hands-on experience with 3D character data structures like SMPL models, kinematic parameters, and hierarchical skeletons.
• A background in physics-based animation or biomechanical motion simulation.
• Know-how in creative software like Blender, Maya, or Unity so you can independently test, visualize, and export your ML outputs.