Lahrech M.

Associate Professor in Electrical Engineering and AI/ML & Embedded Systems

Engineer with 15+ years of hands-on experience, combining applied artificial

intelligence, IoT development, and industrial inspection expertise.

FULL STACK AI — from data to deployment

End-to-end AI solutions: data acquisition and preprocessing, model training

on local GPU (PyTorch, CNNs, QLoRA fine-tuning), through to functional

applications. Recent work includes a fully local RAG pipeline built from

scratch (ChromaDB, Ollama, multilingual embeddings) for technical document

retrieval, and deep learning models for defect detection in industrial

inspection. Currently completing the deployment stack (FastAPI, Docker,

cloud/VPS) to deliver production-ready AI services.

FULL STACK IoT — from sensor to dashboard

Design of connected systems across the entire chain: sensor selection and

electronics (environmental, thermal, optical, GPS modules), embedded firmware

(ESP32, Arduino, MicroPython), wireless transmission (GSM/GPRS, Wi-Fi), cloud

backends (Firebase, REST APIs), and real-time monitoring dashboards. Delivered

projects include a GPS tracking system, a thermal-visible imaging device, and

smart environmental monitoring controllers.

INDUSTRIAL EXPERTISE — NDT, railways, electromechanics

PhD researcher specialized in electromagnetic non-destructive testing (eddy

currents, GMR sensors, 3D-FEM simulation), with applications to railway

infrastructure inspection and composite materials. Additional background in

hands-on electromechanical diagnosis and repair — a deep practical

understanding of real industrial systems.

Researcher's rigor, field engineer's pragmatism: from raw idea to working

prototype. Fluent in English, French, and Arabic.

- Designed and delivered 15+ embedded systems and IoT solutions: GPS trackers (ESP32/SIM800L), environmental monitoring (BME680, CCS811), optical sensing devices, and thermal imaging systems (AMG8833), from hardware selection to firmware and cloud integration (Firebase, REST APIs).

- Developed applied AI/ML solutions: deep learning models for defect detection and inverse problems (PyTorch, CNNs), local RAG pipelines (ChromaDB, Ollama, embeddings), and LLM fine-tuning (QLoRA/Unsloth).

- Built full data workflows in Python: acquisition, preprocessing, model training on GPU, evaluation, and technical documentation.

- Background: university researcher in electrical engineering (electromagnetic NDT, 3D-FEM simulation), bringing rigorous validation methodology to every project.

▸ Teaching electrical engineering, railway traction and industrial maintenance

▸ Supervision of Master's dissertations — 21 dissertations (2014 –2024)

▸ Research and publications in international journals and conferences