Leads the team this role joins
In-house, owns the opening
Put the opening up
From your own Gmail and LinkedIn.
About the role
This is a machine learning role on Mecka AI's computer vision team, hybrid across San Francisco, New York and the Toronto area, for engineers focused on pose estimation, SLAM or 3D reconstruction. It suits engineers who train vision models and then keep them reliable in production. The work turns messy footage of people and objects into accurate 3D data that robotics teams use for training. Mecka builds data infrastructure for robotics and embodied AI.
What you’ll do
- Train vision models from scratch on large multi-view, multi-sensor datasets, adding new losses and architectures where needed.
- Handle difficult footage with motion blur, fast rotation, heavy occlusion and partial views of hands and bodies.
- Measure model error and failure rates on internal video and against public benchmarks.
- Optimize GPU performance, make results reproducible, integrate models into data pipelines and ship releases that can be rolled back.
- Reproduce customer-reported issues, fix them, and prototype quickly on new perception problems.
What they’re looking for
- Strong machine learning fundamentals, with comfort building and training models in Python and PyTorch.
- Experience training large vision models from scratch, beyond running or fine-tuning existing models.
- Depth in at least one computer vision area such as pose estimation, SLAM or 3D reconstruction.
- Experience owning a model or perception component in production after it ships.
- Ability to make progress with noisy real-world data and adapt when priorities shift.
Nice to have
- Hand and body tracking, mesh recovery, human-object interaction and contact estimation.
- SLAM, visual-inertial odometry, learned ego-motion, optical flow, online calibration and metric scale.
- 3D reconstruction and objects, including Gaussian splatting, NeRF and neural 3D object estimation.
What’s on offer
- Pay range of CA$140,000 to CA$200,000 per year.
- Hybrid working across San Francisco, New York and the Toronto area.
Skills
Questions about this role
Is this role hybrid?
Yes, the role is listed as hybrid across San Francisco, New York and the Toronto area.
How much does it pay?
The posted range is CA$140,000 to CA$200,000 per year.
Do I need experience in pose estimation or SLAM?
The posting asks for depth in at least one computer vision area such as pose estimation, SLAM or 3D reconstruction.