Mohamed Emam

Key Skills

Software

CVAT

LabelImg

Label Studio

Roboflow

Top Subject Matter

No subject matter listed

Top Data Types

Image

Text

Video

Top Task Types

Action Recognition

Bounding Box

Classification

Computer Programming Coding

Object Detection

Freelancer Overview

With 7 years of experience in Artificial Intelligence and Machine Learning, I specialize in developing and optimizing high-quality AI training data pipelines, ensuring accuracy and scalability across diverse domains. My expertise spans advanced data labeling for object detection, classification, segmentation, and activity recognition, along with hands-on work in fine-tuning large language models (LLMs) and evaluating prompt responses in both English and Arabic. I bring a deep understanding of how well-structured, contextually rich data fuels AI performance, making me adept at bridging the gap between raw data and production-ready AI systems. Beyond core annotation, I have contributed to building multimodal-RAG systems, intelligent chatbots, and ML models for regression and classification, supported by robust MLOps practices that streamline deployment and monitoring. This blend of technical and applied experience allows me to deliver AI solutions that are not only precise but also adaptable to real-world challenges that setting me apart as a professional who can elevate both data quality and model performance.

Labeling Experience

Cement Factory Operations Monitoring

Label StudioVideoBounding BoxPolyline

This project developed an AI-based Computer Vision System to automate visual monitoring and logistics tracking in a cement manufacturing plant. The system performed multiple tasks, including counting trucks entering and exiting the plant, detecting container fill levels (full or empty), recognizing truck license plates, and monitoring loading/unloading activities. Additional modules tracked weighbridge operations, safety gear compliance for workers, and material spillage detection around loading zones. I contributed to data labeling and validation, creating precise annotations for vehicles, containers, and equipment across varied lighting and dust conditions. The dataset included thousands of hours of CCTV footage processed into labeled frames. A multi-stage QA workflow ensured over 95% annotation accuracy, enabling reliable deep learning models for real-time plant monitoring and operational analytics.

This project developed an AI-based Computer Vision System to automate visual monitoring and logistics tracking in a cement manufacturing plant. The system performed multiple tasks, including counting trucks entering and exiting the plant, detecting container fill levels (full or empty), recognizing truck license plates, and monitoring loading/unloading activities. Additional modules tracked weighbridge operations, safety gear compliance for workers, and material spillage detection around loading zones. I contributed to data labeling and validation, creating precise annotations for vehicles, containers, and equipment across varied lighting and dust conditions. The dataset included thousands of hours of CCTV footage processed into labeled frames. A multi-stage QA workflow ensured over 95% annotation accuracy, enabling reliable deep learning models for real-time plant monitoring and operational analytics.

2024 - 2024

Airport Operations Automation

RoboflowVideoBounding BoxPolygon

This project developed an AI-powered Computer Vision System to optimize airport operations and safety. The system handled multiple tasks, including check-in counter monitoring, luggage counting on belts, passenger flow tracking, activity recognition (e.g., queueing, sitting, restricted area movement), and crowd density analysis. Additional modules detected ground vehicles and monitored staff presence across terminals and runways. I was responsible for data labeling and quality assurance, producing precise bounding boxes and activity annotations across diverse scenes. The dataset contained tens of thousands of images and video frames, capturing varied lighting and camera angles for model robustness. A rigorous QA process, combining peer reviews and automated validation, maintained over 95% annotation accuracy, ensuring reliable training data for real-time airport analytics.

This project developed an AI-powered Computer Vision System to optimize airport operations and safety. The system handled multiple tasks, including check-in counter monitoring, luggage counting on belts, passenger flow tracking, activity recognition (e.g., queueing, sitting, restricted area movement), and crowd density analysis. Additional modules detected ground vehicles and monitored staff presence across terminals and runways. I was responsible for data labeling and quality assurance, producing precise bounding boxes and activity annotations across diverse scenes. The dataset contained tens of thousands of images and video frames, capturing varied lighting and camera angles for model robustness. A rigorous QA process, combining peer reviews and automated validation, maintained over 95% annotation accuracy, ensuring reliable training data for real-time airport analytics.

2024 - 2024

Construction Site Vehicle Monitoring

Label StudioVideoBounding BoxSegmentation

As part of this project, we built an intelligent vehicle monitoring system using advanced computer vision and deep learning techniques to enhance security and logistics management at large construction sites. The system automatically identifies vehicle types (trucks, excavators, mixers, loaders), captures license plates, and detects color and model. It also estimates entry and exit times, measures vehicle speed, and tracks load status to distinguish between loaded and empty trucks. The solution integrated with CCTV infrastructure for real-time monitoring and data analytics, providing operational insights and reducing manual supervision. I was responsible for image annotation, validation, and QA, ensuring high-quality training data from diverse camera angles and lighting conditions. Through continuous QA cycles, the dataset achieved over 96% labeling accuracy, supporting deployment-ready models for site automation.

As part of this project, we built an intelligent vehicle monitoring system using advanced computer vision and deep learning techniques to enhance security and logistics management at large construction sites. The system automatically identifies vehicle types (trucks, excavators, mixers, loaders), captures license plates, and detects color and model. It also estimates entry and exit times, measures vehicle speed, and tracks load status to distinguish between loaded and empty trucks. The solution integrated with CCTV infrastructure for real-time monitoring and data analytics, providing operational insights and reducing manual supervision. I was responsible for image annotation, validation, and QA, ensuring high-quality training data from diverse camera angles and lighting conditions. Through continuous QA cycles, the dataset achieved over 96% labeling accuracy, supporting deployment-ready models for site automation.

2023 - 2023

Personal Safety Equipment Detection

CVATImageBounding Box

Personal Protective Equipment (PPE) Detection Project: This project aimed to develop an AI-driven Personal Protective Equipment (PPE) Detection System to improve worker safety at construction and industrial sites. Using computer vision, the system automatically identified whether workers wore required gear such as helmets, vests, gloves, and boots from images and video feeds. I contributed to data labeling and quality assurance, creating accurate bounding boxes and segmentation masks for PPE and non-PPE items to support training deep learning models. The dataset included over 50,000 images and 300,000 labeled instances, representing varied environments and conditions. To ensure reliability, a multi-level QA process was implemented, including peer review and random sampling, achieving over 95% annotation accuracy. This high-quality dataset enabled the development of robust detection models capable of real-time safety compliance monitoring.

Personal Protective Equipment (PPE) Detection Project: This project aimed to develop an AI-driven Personal Protective Equipment (PPE) Detection System to improve worker safety at construction and industrial sites. Using computer vision, the system automatically identified whether workers wore required gear such as helmets, vests, gloves, and boots from images and video feeds. I contributed to data labeling and quality assurance, creating accurate bounding boxes and segmentation masks for PPE and non-PPE items to support training deep learning models. The dataset included over 50,000 images and 300,000 labeled instances, representing varied environments and conditions. To ensure reliability, a multi-level QA process was implemented, including peer review and random sampling, achieving over 95% annotation accuracy. This high-quality dataset enabled the development of robust detection models capable of real-time safety compliance monitoring.

2023 - 2023

Education

Work History