Professional Portfolio
Technical Projects and Coursework
(Click on each title and/or image for links to repositories or documents)
EIE-4 Generative Networks
Implemented state-of-the-art hyperparameter tuning as Bayesian low-fidelity optimisation on specified tasks in a novel polynomial search space.
Implemented a Variational Auto Encoder (VAE) along with a Generative Adverserial Network (GAN) for image generation use-cases (comparitive analysis).
Implemented U-Net and VGG for segmentation and regression analysis to predict age from brain MRI images with MAE < 6 years.
Built a Turing complete CPU on Intel Quartus Prime using the MU0 architecture with additional features from ARM ISA.
Professional certification programme concluding with a capstone machine learning project involving the prediction of accident severity risk given the weather and road conditions.
An IoT system for an educational multi-player quiz game called 'FQUIZIA' using the DE-10 Lite FPGA boards communicating via Python scripts as client-server architecture.
Google’s Coding Challenge for Bright Network Technology Virtual Workshop 2021 involving implementing various components of a command-line application simulating YouTube in Python.
This challenge can be extended by completing the extra extension parts of the specification and also trying to successfully implement a C++ version.
The objective of this coursework is to model the Open Manipulator X robotic arm, to simulate the model's forward kinematics and analytically solve the inverse kinematics to use it for further control of the robot arm in a desired sequential trajectory of motion.
Final programming assessment of first year of my Programme- a semi guided set of assessed tasks related to developing C++ code to read, extract, analyze and represent passive network circuits and perform operations such as calculating complex impedance and deriving transfer functions of RLC networks.
Developed a web application to remotely command a semi-autonomous rover system consisting of an ESP32 microprocessor, an FPGA vision and Arduino controlled motors.
Developed a CNN model for image classification of aquatic species from a given dataset of fish species using NumPy and PyTorch and fine-tuned the model via transfer learning to improve test accuracy.
EIE-2 C to MIPS Compiler SystemVerilog MIPS CPU
Built a C to MIPS compiler written in C++ using Flex for lexical analysis and Bison for parsing, alongside developing a functional System Verilog MIPS1 CPU (and test-bench).