Ahmed Mohammed
AI/ML Engineer
AI/ML Engineer with 2+ years’ experience in anomaly detection, deep learning, and computer vision, adept at solving challenges like defect detection and occluded object recognition.
Experience
- Conducted research to develop models for detecting defects in industrial inkjet printing
- Achieved 87% accuracy using YOLO with manually annotated defect data
- Explored self-supervised learning techniques and diffusion models
- Designed a framework to classify multi-feature images
- Motor Imagery Classification Project
- Managed EEG data with MNE library and TensorFlow
- Applied LSTM, Bi-LSTM, GRU, and CNN-LSTM models for classification
- Assembled, soldered, and tested PCB boards
- Collaborated with engineers to troubleshoot electrical components
- Gained hands-on experience in electronic systems
Education
Oct 2020 - Present
Master in Artificial Intelligence
Johannes Kepler University Linz, Austria
- Thesis: “Anomaly Detection for Industrial Inkjet Printing Using Machine Learning and Computer Vision”
- Courses: Deep Learning, Computer Vision, Advanced Machine Learning
Feb 2015 - Jan 2018
Bachelor of Mechatronics Engineering
Eastern Mediterranean University Famagusta, Cyprus
- Final Project: Built SCARA robotic system for dynamic object tracking using machine vision
Projects
Anomaly Detection for Inkjet Printing
Developed defect detection models for multi-feature industrial images using YOLO and diffusion models, achieving 87% accuracy by combining traditional computer vision with deep learning techniques.
Computer Vision
YOLO
Diffusion Models
EEG Signal Classification Using Bi-LSTMs
Designed and implemented a Bi-LSTM model for classifying EEG signals into distinct categories, achieving 91% accuracy. Used MNE library and TensorFlow for data processing and model development.
Deep Learning
Bi-LSTM
TensorFlow
Occlusion Removal Using Deep Autoencoder
Developed a deep autoencoder-based solution to enhance object detection accuracy under occlusion. Combined state-of-the-art tracking models with custom detection algorithms.
Computer Vision
Autoencoders
Object Tracking
SCARA Robotic System
Designed and implemented a SCARA robotic system capable of real-time object tracking using computer vision. Integrated motion control systems with visual servoing for precise manipulation.
Robotics
Machine Vision
Control Systems
Contact
Feel free to contact me for any questions or opportunities.