Mohammad Nemer
AI Researcher & IoT Specialist | PhD in Computer Science
AI Researcher & IoT Specialist | PhD in Computer Science
I'm an AI researcher and IoT specialist with over 3 years of experience bridging the gap between theoretical machine learning models and practical embedded systems. My work focuses on developing efficient AI algorithms for resource-constrained IoT devices, with applications in smart manufacture and industrial automation. My PhD research at Université Marie & Louis Pasteur explored novel techniques for distributed machine learning across IoT networks. I've collaborated with top-tier companies including BMW.
I'm passionate about developing AI solutions that solve real-world problems, particularly those that improve sustainability, accessibility, and quality of life.
Designed and deployed an advanced edge AI system that delivers real-time anomaly detection and automated reporting, enabling immediate insights and proactive response by processing data directly at the source for reduced latency and enhanced operational efficienc.
Outcome: Predicts faults before happening
DemoUsed a YOLO-based model to automatically detect defects and anomalies in products during production stages, enabling real-time quality control and faster response to issues on the production line
Outcome: This work helped catching defects faster, improved product quality, and made the production process more efficient
DemoCreated a comprehensive semantic interoperability flow that transforms heterogeneous data formats into unified, machine-readable knowledge graphs using semantic web standards to demonstrate practical data integration and discovery
Outcome:Unified, machine-readable knowledge graphs enabling seamless semantic data integration and discovery
DemoTransitioned the model architecture from MAPFAST to EfficientNet, implementing modifications to the last two layers to optimize performance.
Outcome:Increased the accuracy from 72% to 85%
DemoExplored novel techniques for deploying sophisticated AI models across manufacturing environment.
Full Thesis Coming soonConducted extensive testing of convolutional neural network (CNN) models specifically designed for tabular data by implementing and experimenting with one-dimensional (1D) convolutional layers
Lightweight feature-based priority sampling technique for Industrial IoT multivariate time series that optimizes data transmission reduction and classification accuracy by selectively sampling high-entropy features, demonstrating enhanced ResNet model accuracy (92.3% vs 79.1% baseline) with 40% reduced processing time through noise reduction and efficient data prioritization
Proposes using AI-driven Named Entity Recognition (NER) models to improve semantic interoperability in IoT systems
Conducted extensive testing of time series classification by implementing and experimenting common transformer architectures
The paper introduces AI4PM, a framework that leverages distributed artificial intelligence to enable intelligent, autonomous behavior and self-reconfiguration in large-scale programmable matter systems composed of modular robots
Awarded for my academic performance and granted a scholarship to study in France
Led lab sessions on IoT protocol design, developed course materials on edge AI deployment, and mentored student projects
Conducted a 2 day workshop on deploying neural networks to resource-constrained IoT devices
Mentored a master's students and 2 undergraduate research assistants on machine learning projects
I'm always interested in discussing new opportunities, collaborations, or simply connecting with fellow researchers and professionals. Feel free to reach out!