Dissertations
Graph Neural Networks for Accelerated Genome Assembly
My postgraduate dissertation project using techniques from Geometric Deep Learning to resolve the true genome sequence in genome assembly graphs built from erroneous reads.
Physically based CUDA-Accelerated Path Tracer
My undergraduate dissertation project utilizing the Cook-Torrance Microfacet Model, acceleration data structures, several BRDF importance sampling techniques, and deep learning denoising to achieve photo-realistic rendering on NVIDIA GPUs.
Other projects
Reinforcement Learning for Hydrophobic-Polar Protein Folding
Achieved state of the art results using Double Deep Q-Networks, Mamba Selective State Space Models, and Graph Attention Networks on the HP model for protein folding.
Natural Language Parser Comparison
Evaluated and compared the performance of two natural language constituency parsers from quantitative and linguistic standpoints.
Computer Architecture Essays
A collection containing my writings from computer architecture papers discussing trends, state-of-the-art microprocessor & memory design, hardware reliability, specification, verification & test, security, and hardware accelerators for machine learning.
Optimizing a spark-ignition engine
Built a thermodynamically accurate engine simulation. Gaussian Processes and Bayesian Optimization was then used to optimize the engine design to maximize thermodynamic efficiency.
SLR(1) parsing table generator
Simple Left-to-Right Rightmost-derivation 1-symbol look-ahead parsing table generator for deterministic context-free languages.