Mohamed Moustafa

Key Skills

Software

Internal/Proprietary Tooling

Top Subject Matter

No subject matter listed

Top Data Types

Computer Code Programming

Image

Text

Top Task Types

Bounding Box

Classification

Computer Programming Coding

Evaluation Rating

Object Detection

Freelancer Overview

I have extensive experience in building and managing AI training data pipelines, with a focus on ensuring data quality, consistency, and scalability for machine learning applications. At Robert Bosch GmbH, I developed automated ETL pipelines for hybrid data integration (real and synthetic), which significantly improved object detection performance in safety-critical scenarios. I also designed and implemented a synthetic data quality assessment framework, helping to validate datasets before training and ensuring robust model generalization. In addition, I worked on data labeling and ontology automation, creating Python-based APIs and a Docs-as-Code system with Sphinx-Needs that reduced documentation and labeling errors by 40%. This streamlined collaboration across teams and improved traceability of datasets. My experience extends to defining data quality standards, detecting dataset gaps, and developing validation scripts to ensure consistency—skills that are particularly valuable in high-stakes domains such as autonomous driving and driver assistance systems.

Labeling Experience

Education

Work History