Hardware/Software engineering project | 6 months R&D at Medivietech (MedTech Startup)
Clinical validation: 50+ patients | 400+ hours of data | 99.2% system uptime
- 📐 System Architecture — Hardware/software architecture overview
- 🔧 Technical Challenges — Engineering problems solved
- 🏥 Clinical Validation — Hospital deployment results (50+ patients)
- 🔌 Hardware Guide — Component specifications & PCB design
Professional CAD renders of PULSAR prototypes |
Functional prototypes PULSAR-007 & PULSAR-009 |
Deported circular sensor via FPC ribbon cable |
Internal integration: LiPo battery + PCB |
PULSAR is a medical-grade wearable device for continuous physiological monitoring in hospital environments. Unlike consumer wearables (Fitbit, WHOOP), PULSAR achieves clinical-level precision validated against professional medical equipment.
My role: Complete hardware/software development from legacy prototype debugging to clinical deployment.
| Metric | Result |
|---|---|
| Clinical Validation | 50+ patients @ Clinique Hartmann (Neuilly-sur-Seine) |
| Data Collected | 400+ hours of medical-grade recordings |
| System Reliability | 99.2% uptime during 3-month deployment |
| Data Integrity | 0% data loss (FIFO optimization) |
| Prototypes Built | 15+ functional units |
| Battery Life | 8-10h (ESP32 prototypes) → 3-5 days target (Nordic production) |
- ✅ Multi-channel optical sensor integration (PPG 4-channel, 100 Hz sampling)
- ✅ Custom PCB design with deported circular sensor (ergonomic wearability)
- ✅ Power optimization (fuel gauge monitoring, LiPo charging circuit)
- ✅ Mechanical CAD design for 3D-printed enclosures (15+ iterations)
- ✅ Sensor fusion (accelerometer + PPG for motion artifact rejection)
Circular deported optical sensor (Ø10mm) connected via flexible FPC ribbon cable
- ✅ Real-time firmware for dual-core ESP32-S3 (240 MHz)
- ✅ Zero-loss data pipeline (resolved critical FIFO overflow bug)
- ✅ Dual-mode architecture (WiFi + AWS cloud / Standalone SD card recording)
- ✅ Protocol implementation (I2C @ 400kHz, SPI, FAT32 file system)
- ✅ Memory optimization (PSRAM management for continuous acquisition)
- ✅ PPG waveform analysis (multi-spectral: Red, IR, Green, Blue channels)
- ✅ Data encoding (JSON chunking + Base64 for cloud upload)
- ✅ Clinical-grade accuracy validated against hospital equipment
Photoplethysmography (PPG) measurement principle |
Typical PPG signal with cardiac features |
Location: Clinique Hartmann (Neuilly-sur-Seine, France)
Supervising Physician: Dr. Lee (Intensive Care Unit)
Duration: 3 months continuous testing
Patient Demographics: 50+ patients (diverse age groups, pathologies)
- ✅ Medical precision validated against professional oximeters
- ✅ 99.2% system uptime (only 2 units had minor issues over 90 days)
- ✅ User comfort confirmed by medical staff and patients
- ✅ Zero safety incidents throughout deployment period
Inherited Problem: Original firmware discarded 29 out of 30 samples
Root Cause: Incorrect FIFO read logic + buffer overflow
Solution: Complete redesign of data acquisition pipeline
Result: 0% data loss on all 15 deployed units
Problem: SD card + Accelerometer sharing SPI bus caused system crashes
Solution: Exclusive SPI access with proper timing guards
Result: Stable 8-10h continuous recording sessions
Problem: Square PCB uncomfortable for wrist wear
Solution: Circular deported sensor (Ø10mm) via flexible FPC cable
Result: Positive comfort feedback from 50+ patients
Microcontrollers:
- ESP32-S3 (R&D prototyping: 240 MHz dual-core, 8MB PSRAM)
- Nordic nRF5340 (Production target: ARM Cortex-M33, ultra-low power)
Sensors & Components:
- Multi-spectral optical sensor (PPG 4-channel @ 100 Hz)
- 3-axis accelerometer (motion detection)
- Fuel gauge IC (battery monitoring)
- MicroSD card (FAT32 storage)
Connectivity:
- WiFi 802.11n (AWS S3 upload pipeline)
- Bluetooth 5.0 LE (future mobile app support)
Development Tools:
- Firmware: C/C++, Arduino/ESP-IDF, VS Code, Git
- PCB Design: Altium , EasyEDA Pro
- Mechanical CAD: SolidWorks, Fusion 360
- Cloud: AWS S3, AWS Lambda
- Analysis: Python (data extraction, FFT analysis)
Spectral analysis of 4-channel PPG data (cardiac frequency band highlighted)
PULSAR/
├── README.md # Portfolio presentation (this file)
├── LICENSE.md # Confidentiality notice
├── docs/ # Technical documentation
│ ├── architecture.md # System architecture overview
│ ├── challenges.md # Technical challenges solved
│ └── validation.md # Clinical validation results
├── hardware/ # Hardware specifications
│ ├── README.md # Hardware guide
│ └── component-list.md # Bill of Materials
└── images/ # Visual documentation
├── prototypes/ # Physical device photos
├── hardware/ # PCB and component photos
├── cad/ # CAD renders
└── architecture/ # Technical diagrams
Note: Hardware schematics, firmware source code, and detailed system architectures remain confidential per Medivietech company policy. Documentation in /docs/ provides high-level technical context without revealing proprietary implementations.
- Embedded Systems: Firmware development, RTOS concepts, memory optimization
- Hardware Design: PCB design, sensor integration, power management
- Signal Processing: PPG analysis, FFT, artifact rejection algorithms
- CAD Engineering: Mechanical design for wearable devices
- IoT Architecture: WiFi connectivity, cloud integration (AWS), data pipelines
- Medical Devices: Clinical validation methodology, hospital deployment
- R&D Project Management: Led project from debugging to clinical validation
- Cross-functional Collaboration: Worked with medical staff, regulatory teams
- Technical Documentation: Created multilingual documentation for international partners
- Problem-Solving: Diagnosed and resolved critical system bugs under time pressure
- Startup Environment: Operated autonomously with minimal supervision
This repository showcases my engineering work during my internship at Medivietech (AGORANOV Paris).
Intellectual property belongs to Medivietech.
✅ Authorized: Portfolio review for recruitment/freelance evaluation
❌ Not Authorized: Commercial use, code reproduction, derivative products
For commercial inquiries: contact@medivietech.com
Tom HUYGHE
Mechatronics Engineer | Embedded Systems Specialist | MedTech Innovator
🎓 ESME SUDRIA — Engineering Degree (Mechatronics & Embedded Systems)
🏢 Currently: Freelance embedded consultant + Open to full-time opportunities
💼 Interests: MedTech, IoT, Hardware, Wearables, Signal Processing
📧 Email: [tom.huyghe@orange.com]
💼 LinkedIn: [linkedin.com/in/tom-huyghe]
🌐 Portfolio: [tomhuyghe.dev]
💰 Freelance Rate: 350€/day
- Embedded firmware (ESP32, Nordic nRF, STM32)
- PCB design & rapid prototyping
- Physiological signal processing (PPG, ECG, IMU)
- Medical device development & validation
- IoT cloud architectures
- Python tooling & data analysis
Company: Medivietech (MedTech Startup)
Incubator: AGORANOV Paris (Leading French Deep-Tech Incubator)
Role: Hardware/Software Engineering Intern
Duration: 6 months (April - October 2024)
Team: CEO (Neil Benhamou), CTO (Thomas Baret - AI/Data),
International Partnership: EMBRILL (India) for industrialization
Special thanks to:
- Neil Benhamou (CEO, Medivietech) — For entrusting me with technical autonomy
- Thomas Baret (CTO, Medivietech) — For AI/Data collaboration
- Dr. Lee (Clinique Hartmann) — For clinical validation support
- AGORANOV — For the inspiring startup environment
- EMBRILL Team (India) — For industrialization partnership
⚡ Engineered with passion for MedTech innovation ⚡
Transforming engineering skills into medical devices that improve lives
Last Updated: January 2025
Project Status: ✅ Clinical validation completed | ✅ Production migration ongoing (Nordic nRF5340)