University of Nottingham
MSci Computer Science
- Achieved First Class Honours (1:1).
- Awarded Best Undergraduate Research Project.
Core platforms and frameworks for Wealth Services.
Conducted and led a novel research project exploring the use of synthetic datasets generated by 3D Gaussian Splatting, 2D Diffusion, and other generative AI methods to improve performance and image segmentation fidelity in plant phenotypingb and disease segmentation tasks.
Served as Fullstack Engineer and Team Lead for a VR compatibility layer integrating retro text-based adventure games, overseeing a team of 8 developers and ensuring on-time delivery through agile methodologies, SOLID principles, and structured stand-ups. Additionally spearheaded the development of an API system and emulator backend using C# and C, and awarded "Best Showcase" for delivering a technically ambitious, standout user experience.
Supervised 21 Junior Assistants whilst providing front-line technical support to 44,000+ students and faculty in a timely and effective manner. Assisted in troubleshooting hardware, software, and network-related issues by utilising Microsoft Azure Entra ID, networking management, and customer service skills to properly diagnose issues and escalate faults to ensure customer satisfaction.
Citrus is a 3D game engine that aims to deliver high performance without compromising graphical quality.
A custom graphics/rendering engine built in OpenGL that was developed as an aide whilst learning OpenGL. It includes Volumetric Lighting, IBL Lighting, PBR, HDR, Ambient Occlusion, Compute Shader support, Deferred Rendering, etc.
An optimisation framework for constrained optimisation of hetereogeneous rainbow tables implemented using Python and various search algorithms, reinforcing prior research and identifying further performance enhancements in TMTO based attacks.
A novel research project exploring the use of synthetic datasets generated by 3D Gaussian Splatting and 2D Diffusion to improve performance and image segmentation fidelity in plant phenotyping segmentation tasks.
Modular ROS2 autonomy stack for exploration and navigation using frontier clustering and A*-based safe-goal synthesis with radial fallback. Trained a YOLOv8s detector on blended real+synthetic data with Blender-based domain randomisation.
A semantic image segmentation pipeline custom build to address data scarity scenarios in precision agricultural contexts. Implemented methods consist of a custom composite loss function, vegetation indices, an aggressive image augmentation stack, and sliding window inference on DeepLabV3+. Achieves 0.83 mIoU and 97% accuracy.
A VR compatibility/translation layer for retro text-based games that allows for a unique user experience and retelling of older games, without the need for porting existing codebases to new systems.
An intelligent AI-powered file sorting application using Python developed during a 24-hour hackathon to tackle disorganised file systems by automating file categorisation and sorting with precision
An toy imperative programming language developed using functional programming via Haskell.
A simple minecraft plugin to relay messages between minecraft and discord servers and allow for account linking and whitelisting capabilities.
An operating system process scheduling simulator built using C and POSIX APIs. Implemented various concurrency methods and OS systems such as burst/round robin scheduling, mutual exclusion and semaphores, load balancing, interrupt handling, etc.
A heuristic based problem optimiser for the variant of a dynamic TSP problem built using Java and HyFlex HyperHeuristic framework.
You're looking at it :)