Professional Summary
Embedded Systems Engineer with 4+ years of experience, developing custom Linux kernels, RTOS, and implementing device drivers and applications for STM32 and NXP platforms. As a PhD Researcher, working on advanced and edge AI for time-series anomaly detection, focusing on signal processing. Curious and self-driven who enjoys mastering new technologies quickly and continuously expanding technical expertise.
Professional Experience
- Developed 3+ anomaly detection algorithms using deep learning in IMU/GNSS/5G real-time processing
- Enhanced step detection by 23% using Embedded model in smarphone on 10,000+ data samples across 5 environments
- Analyzed 800+ minutes of GNSS signal anomalies, identifying jamming and spoofing
- Improved 5G signal detection LOS/NLOS in indoor environments
- Developed 3+ custom Linux images using Yocto on 3 hardware platforms (STM32MP1, i.MX93)
- Implemented 10+ device drivers for peripherals (GPIO, I2C, SPI, UART, LoRaWAN, Zigbee)
- Developed a C++ embedded Linux application that orchestrates all device drivers and integrates with Pure Data (Pd) to deliver the full interactive behavior of the Muzziball system.
- Designed distributed measurement systems for remote monitoring using LabVIEW and Python
- Conducted 50+ lab sessions for 150+ students
- Working on Antenna Design (HFSS/CST), Microwave Engineering, and RF Systems courses
Key Projects
Developing custom Linux image with Yocto for STM32MP1, integrating IMU, Audio, BLE, WiFi, and charging system. Implementing embedded C++ applications for real-time data acquisition.
Custom Linux image with Yocto for NXP i.MX93, integrating 5G, GNSS, IMU, and CAN bus for vehicle tracking. Developed all device drivers for peripherals.
Developed lightweight deep learning-based anomaly detection algorithms for pedestrian dead reckoning (PDR) using IMU data.
Enhanced step detection accuracy by 23% through unsupervised anomaly filtering on 10,000+ samples across 5 environments.
GitHub
Designed and implemented deep learning models for Line-of-Sight (LOS) and Non-Line-of-Sight (NLOS) detection in 5G indoor positioning systems.
GitHub
Developed deep learning models for breast cancer detection using mammography images from the RSNA screening dataset. Implemented CNN architectures for tumor classification and localization, applying data augmentation and transfer learning techniques to improve diagnostic accuracy.
View on GitHub
Built NLP deep learning models for detecting contradictory statements in multilingual text pairs. Utilized transformer-based architectures (BERT) for natural language inference (NLI), achieving robust classification of entailment, neutral, and contradiction relationships.
View on GitHub
Built wireless sensor network for crack detection in heritage buildings (Hardware + Firmware + Cloud Dashboard).
IEEE MetroLivEnv 2023 Paper
Publications & Research
Education
Technical Skills
Languages
Actively developing French skills and very open to working in a French-speaking environment.