Teams at Mecka AI

Lead and scale end-to-end data annotation operations: set quality standards, build and manage an in-house annotation team and vendors, translate research into annotation guidelines, implement QA systems, drive AI-assisted labeling, partner with ML/CV and product, and own metrics and reporting.

Design and train state-of-the-art 3D hand pose, dense mesh, and articulation models from scratch. Scale multi-view and temporal architectures on multi-GPU clusters, innovate losses enforcing biomechanics and temporal/physical plausibility, and build robust egocentric hand-object interaction models. Rapidly prototype solutions for tactile-visual fusion, action segmentation, and sensor integration, and integrate tracking outputs into physics-aware pipelines for downstream robotic control.

Own, maintain, and improve SLAM, VIO, and SfM pipelines; perform rigorous multi-camera and IMU calibration; ensure cross-device spatial computing robustness; build debugging and visualization tools for images, point clouds, and trajectories; support company-wide computer vision, data validation, and labeling workflows; collaborate closely with hardware teams for sensor calibration and on-site testing.