Subashini Manickam

Embedded Systems Engineer with 2.8 years in IT (automation, support) and a Master’s in Embedded and Real-Time Systems. Skilled in Embedded C, AUTOSAR, CAN, SPI, I2C, RTOS, and MATLAB/Simulink, with hands-on ADAS and driver monitoring projects. Open to embedded software development, testing, and integration roles.

Real-time ADAS-enhanced deep learning framework for detecting driver's drowsiness and various emotional states| Python, OpenCV, TensorFlow, ResNet50, Jupyter Notebook | 06/06/24 - 1/12/24

• Designed a ResNet50-based deep learning model for detecting driver drowsiness, aggression, and emotional states using camera inputs, achieved over 97% detection accuracy, was awarded 1st prize at Research Conclave '25, and was selected for Taylor & Francis publication (pending)

Driver safety assistance system based on the Smith System using Raspberry Pi 4| Python, OpenCV, Raspberry Pi 4| 01/01/25 to 05/05/25

• Developed a real-time drowsiness and pedestrian detection system, implementing a few Smith System driving principles, and integrating image processing algorithms for hazard detection, improving reaction time by approximately 25%

Model-based development project | MATLAB/Simulink | 01/01/24 to 04/04/24

• Built Simulink models for embedded control applications, reducing design iteration time by 15% via simulation-driven testing

I2C and SPI communication implementation | Arduino Mega | 09/09/23 to 10/12/23

• Configured master-slave communication between Arduino boards using I2C/SPI, validated performance under varied conditions

Udemy course - From CAN to Real-Time Systems

Secured first place for presenting my journal paper titled Real-Time ADAS Enhanced Deep Learning Framework for Detecting Driver's Drowsiness and Various Emotional States at Research Conclave '25, organized by PSG College of Technology, among 368 participants

Real-Time ADAS-Enhanced Deep Learning Framework detecting Driver's Drowsiness and Various Emotional States, 1st Prize, Research Conclave'25; selected for Taylor & Francis publication (pending).