Mecka AI
Research Scientist – 3D Environment Reconstruction
$200K–$230K
Toronto GTA +1 · Hybrid · Full-time · Posted Oct 1, 2026
- Location
- Toronto GTA
New York
About the job
Architect and train 3D reconstruction pipelines that turn real-world multimodal data into photorealistic digital environments for robotics and embodied AI, while optimizing those environments for downstream use and client viewing.
What you'll do
- Design, train, and scale neural reconstruction models for dense 3D geometry across multi-GPU clusters.
- Build accurate digital twins that support robotics perception, world models, and sim-to-real training.
- Optimize and package 3D scenes for interactive client viewing, and prototype perception research including dynamic-actor isolation and sensor integration.
What they ask for
- Deep expertise in 3D Computer Vision, Neural Rendering, multi-view geometry, and spatial modeling.
- Proven experience training and scaling large-scale 3D vision models from scratch using PyTorch and multi-terabyte datasets.
- Experience writing custom CUDA kernels for 3D operations, ray marching, or rasterization.
Nice to have
- First-author publications at CVPR, ICCV, or ECCV focused on 3D deep learning, neural rendering, or spatial transformers.
- Public projects deploying 3D assets to WebGL or custom rendering engines, or optimizing splats for interactive viewing.
Pay
CA$200K – CA$230K (base salary)
Tools
3D Gaussian Splatting, NeRFs, PyTorch, CUDA, WebGL
AI focus
Develop neural 3D reconstruction and perception systems for embodied AI and robotics, including sim-to-real integration.
Model-written summary · the original posting is authoritative