Mahdi Mnif

• Deployed Edge AI (Arduino, ESP32, Pi, Coral, MAX78000), cutting power by 52% and halving inference time. • Engineered a 66μs latency MAX78000 solution, achieving 89.4% accuracy for bioimpedance gesture classification. • Built an optimization pipeline reducing model complexity by 94.4%. • Optimized Spiking Neural Networks (SNN) for real-time edge inference. • Mentored Master's researchers and led embedded ML and IoT workshops.

• Executed custom RTL for SNN hardware and FPGA-based SoCs during an international neuromorphic computing research stay. • Compressed SNN architectures by 17x for Edge AI, driving projected energy efficiency gains.

• Developed a 35mm, 4-layer PCB wireless sensor node with physiological and inertial sensors for WBAN health monitoring. • Implemented an RF transmission system using RSSI for real-time patient movement tracking. • Integrated continuous sensor data into Firebase with a web-based visualization dashboard.

Joint PhD candidate (TU Chemnitz & ENET'com); awarded the DAAD Research Grant (2023-2025). Researched TinyML and hardware-aware ML, optimizing CNNs and SNNs for resource-constrained edge deployment. Applied optimized edge models to real-time hand gesture recognition using complex bioimpedance data. Assisted with logistics and registration for international conferences (CIVEMSA, IWIS, ROSE, I2MTC) and the DAAD STEP project.

Note: Equivalent to M.Sc. in Engineering. Active member of the IEEE ENET'COM Student Branch (11/2018 - 03/2021). Supported event logistics for technical and social events, including the IEEEXtreme 12.0 hackathon, and taught programming fundamentals to young pupils as an Hour of Code instructor.

Intensive two-year preparation program in advanced mathematics, physics, and chemistry leading up to the national competitive entrance exams for engineering schools.