Key Skills

Software

Other

Top Subject Matter

No subject matter listed

Top Data Types

Image

Text

Video

Top Task Types

Classification

Data Collection

Object Detection

Freelancer Overview

With extensive experience in AI research, I specialize in data labeling and training data preparation for multimodal datasets, including visual, inertial, and time-series data. My work in human activity recognition (RGB/thermal/IMU), Arabic NLP (social media, lyrics), and predictive maintenance (vision-based vibration analysis) required meticulous annotation, feature engineering, and validation. Projects like VCOACH (virtual coaching) and MM-DOS (workout datasets) involved labeling complex motion patterns and extracting frequency features. I’ve also developed novel descriptors for IMU-based activity recognition and benchmarked deep learning models against traditional methods. My expertise in contactless sensing and real-world applications sets me apart. I’ve led STDF/JICA-funded projects requiring robust labeling for harsh environments (e.g., rotating machinery fault detection using Eulerian motion magnification). In Arabic NLP, I curated and labeled diverse corpora (e.g., Amina news dataset), while my quantum ML work involved hybrid data preprocessing. I’ve supervised 12 PhD and 7 MSc theses on data-centric AI, emphasizing reproducible labeling and adversarial robustness. Key strengths: - Multimodal data fusion (video/IMU/audio) with domain-specific labeling. - Industrial AI applications (predictive maintenance, crowd analytics). - Leadership in large-scale projects with auditable annotation pipelines. - Teaching graduate-level ML/data-centric courses

Labeling Experience

Education

Work History