My projects

Here are some of the projects that I worked on (in reverse chronological order):

1. Pupil as an indication for neuropathic pain – under submission

Objective: Develop an algorithm to track and measure pupil size variation over time.

> Achieved an accuracy of over 90%.

> Discovered that pupil dilates 27.7±5% more under pain.

Technologies used: Python, Scikit-learn, TensorFlow Lite, NumPy, Google Colab, Jupyter Notebook, Raspberry Pi 4 model b.

2. Hardware-aware Automated Architecture Search for Brain-inspired Hyperdimensional Computing – under submission

Objective: To find the right model size based on the hardware specifications and achieve the highest performance.

Improved ROC-AUC score by 4.79% and reduced model size by 106.35 times.

Technologies used: Python, Scikit-learn, TensorFlow Lite, NumPy, Google Colab, Jupyter Notebook, Raspberry Pi 4 model b.

3. Deploying Hyperdimensional Computing (HDC) for Drug discovery tasks on Edge Devices

Objective: To deploy a HDC model (Mole-HD) on an edge device to achieve highest energy efficiency.

Reduced memory usage by 64.38% and increased energy efficiency by 62.44 times as compared to NVIDIA Tesla K80.

Technologies used: Python, C, Scikit-learn, MicroPython, OpenMV, TensorFlow Lite, TensorFlow, NumPy, Google Colab, Jupyter Notebook, Raspberry Pi 4 model b.

4. Fire & Smoke Detection Model for Real-time IoT Applications

Objective: To develop and deploy a smart Fire and Smoke detection IoT system to assure reliable safe homes.

Every part of the system is automated with a 2 way communication from the cloud (fire fighters) and home system.

Technologies used: Arduino NANO 33 IoT, Arduino web editor, Arduino IoT Cloud, MQ-2 Smoke sensor, LM393 light sensor, BN- 220 GPS module, 0.96” OLED, DHT22 Temperature and Humidity sensor.

5. image compression using dct

Objective: To compress an image with minimum distortion using Discrete cosine transform (DCT), Single value decomposition (SVD) using MATLAB.

Technologies used: MATLAB, SVD, DCT

6. Health monitoring using android application

Objective: To develop a system that monitors and records the health condition of disabled persons virtually and transmits the data to health care professionals for appropriate feedback.

Technologies used: C, Code Blocks, Arduino IDE, Arduino UNO, MAX30100 Heart Rate Sensor, LM35 Temperature
Sensor, HC-05 Bluetooth Module, Mobile Phone.

7.Walker – the walking robot

Objective: To develop an autonomous walking robot (insect like). It has the ability to walk, sit, and avoid obstacles without any human intervention.

Technologies used: C, Code Blocks, Arduino IDE, Arduino UNO, SG90S Mini Servo Motor, HC-SRO4 Ultrasonic
Sensors, Photodiodes, Power Bank, Resistors.

8. automatic water dispenser

Objective: To develop a vending machine that dispenses the water based on the coin inserted.

Technologies used: C, Code Blocks, Arduino IDE, Arduino UNO