Research Scientist (3D Human Pose)

Location: On-site in Toronto GTA or New York

Salary: $150K–$200K total wage range, with compensation varying by hiring location and potentially including base salary, incentives, equity, and benefits.

Company

hirify.global builds data infrastructure for robotics and embodied AI, including systems for data capture, labeling, and hardware-enabled workflows used to train and validate humanoid and embodied AI systems.

What you will do

• Design, implement, and train proprietary models for 3D human pose estimation, full-body mesh recovery, and kinematic tracking.
• Scale multi-view and temporal architectures across multi-GPU clusters and massive multimodal human-motion datasets.
• Develop loss functions and model architectures incorporating biomechanical constraints, temporal smoothness, postural balance, and physical plausibility.
• Build human-scene interaction models for severe occlusion, motion blur, multi-person crowding, and complex environments.
• Prototype perception models for action segmentation, intent prediction, and new hardware sensor integrations.
• Integrate tracking outputs into physics-aware pipelines that reason about contact surfaces, gravity, and momentum for robotic control and locomotion.

Requirements

• Deep expertise in deep learning, 3D computer vision, articulated tracking, and human body pose estimation.
• Experience training large-scale vision models from scratch rather than only running inference or fine-tuning existing checkpoints.
• Strong theoretical and practical understanding of parametric human body models such as SMPL, SMPL-X, GHUM, MHR, or SOMA-X, as well as inverse kinematics and dense mesh estimation.
• Mastery of PyTorch and deep learning scaling frameworks.
• Experience curating and handling massive, multi-terabyte image and video datasets.
• Ability to work on-site in Toronto GTA or New York.

Nice to have

• First-author publications at CVPR, ICCV, ECCV, or NeurIPS focused on 3D human pose tracking, human-scene interaction, motion capture, or human mesh recovery.
• Experience with large human-motion datasets such as AMASS, Human3.6M, EgoBody, or PROX.

Culture & Benefits

• Pure R&D and ownership of new models without maintaining legacy systems.
• Access to proprietary, high-quality spatial and temporal ground-truth human-motion data.
• Compute resources for large-scale model development and state-of-the-art research.
• Work in a fast-paced environment where priorities can shift with new research and hardware capabilities.
• Potential equity, health and wellness benefits, and other company programs.
• Inclusive hiring practices and interview accommodations when needed.