Teams at Mecka AI

Build and maintain an internal computer vision annotation platform, including interactive workflows, dashboards, data pipelines, and quality measurement tools. Develop customer-facing ML services for PII blurring and model-assisted labeling. Deploy and optimize detection and segmentation models using PyTorch, TensorRT, and ONNX, and expose them through high-performance concurrent APIs. The role requires strong frontend engineering, production Python, applied ML deployment, video processing, and human-in-the-loop workflow expertise.

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.