SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Mecka AI is building the data infrastructure layer for robotics and embodied AI. The company designs and operates global systems for data capture, data labeling, and hardware-enabled workflows used by leading AI labs and robotics companies to train and validate humanoid and embodied AI systems.
You will own the classical computer vision side of the stack, transforming raw per-frame model predictions into clean, rigid, smooth, correctly-calibrated spatial data. You'll partner closely with research teams to understand deep learning models without training them yourself.
Your immediate focus is the 3D hands pipeline, where classical CV work is most critical today. From there, you'll serve as the organization's classical-CV specialist, brought in for geometry, filtering, and pipeline problems across the company.
Key responsibilities:
- Own classical 3D geometry: multi-view geometry, triangulation, PnP, camera calibration, undistortion, and reprojection to place reconstructions correctly in space.
- Design and tune temporal filters (low-pass, Kalman-family, interpolation) that remove jitter while preserving real motion.
- Run inverse kinematics and enforce physical constraints (bone-length rigidity) to keep reconstructed skeletons valid; build QA checks that gate delivery.
- Turn model output into clean delivered data: per-frame results, reports, QA metrics, and pipeline operations at scale.
- Build visualizers and dashboards to inspect trajectories, overlays, and failure cases.
- Be the go-to engineer for classical computer-vision problems across the pipeline.
Requirements:
Must-have:
- 5+ years of professional software engineering and computer vision experience.
- Strong software engineering fundamentals: clean, efficient, production Python (C++ a plus).
- Solid multi-view geometry: triangulation, PnP, camera calibration, reprojection, coordinate transforms.
- Temporal filtering and signal processing: low-pass and Kalman-family filters, interpolation, smoothness/fidelity trade-offs.
- Inverse kinematics and rigid-body or bone-length constraints on articulated structures.
- Enough ML knowledge to work with deep models (pose and mesh regressors), understanding their outputs and failure modes, without training them.
- Data-quality mindset: comfortable with noisy real-world data and building QA that catches problems before customers.
Nice-to-have:
- Human pose or mesh reconstruction (hand or body), keypoint and mesh pipelines.
- Depth in Bayesian filtering or state estimation; experience with optimization frameworks (Ceres, GTSAM).
- CV tooling: trajectory visualizers, overlay or QA dashboards.
- Stereo, multi-view, or ego-exo capture experience.
- Light temporal ML (small networks for smoothing or gap-filling).
