Welcome to My Portfolio

PhD student in Mechanical Engineering | Computational Physics & Machine Learning

Background

About Me

I am a PhD student in Mechanical Engineering at Iowa State University specializing in computational physics, machine learning, and GPU-accelerated simulation methods. My research bridges classical physics-based methods such as finite element, material point, and discrete element methods with advanced soil models and machine learning.

My other research interests include: Autonomous vehicles, vehicle dynamics, nonlinear control systems, random signal processing, Bayesian statistics, optimisation and probabilistic model inversion of complex physics-based models.

I ❤️

Science 🧪 | Engineering 🚀 | Adventure 🚁 | Travel 🌍 | Fitness 💪 | Rock Climbing 🧗 | Snowboarding 🏂 | Surfing 🌊

About me

Research Focus 🧪

Numerical Modeling

Finite element modeling with density-dependent NorSand constitutive models for accurate terrain response prediction.

Experimental Investigation

Comprehensive studies on factors influencing bevameter terrain response, including plate sinkage and shear tests.

Terrain-Vehicle Systems

Research into complex vehicle–terrain interactions, advancing mobility modeling and off-road performance prediction.

Education 🎓

Degree University Years GPA
Doctorate of Mechanical Engineering Iowa State University (USA) Jan 2023 – Dec 2026 3.82 / 4.0
Master of Mechanical Engineering University of Pretoria (South Africa) Jul 2020 – Dec 2022 4.0 / 4.0

Graduate Courses 📚

  • Machine Learning for Cyber-Physical Systems (A)
  • Machine Vision (A)
  • Vehicle Dynamics (A-)
  • State-Space Control Systems (A)
  • Continuum Mechanics (A-)
  • Advanced Finite Element Methods (B+)
  • Modeling of Soil-Crop-Machine Systems (A)
  • Numerical Methods in Geotechnical Engineering (A)
  • Mechanical Vibration (A)
  • Fatigue (A)
  • Bayesian Statistics (A)
  • Soil Mechanics (B)
  • Statistical Methods and Experimental Design (A-)
  • Deep Learning (A)
  • Internet-Based Methods (A)
  • Scientific Machine Learning (A)

Published Research 🏛️

NorSand FEM simulation of Bevameter Tests (2025)

Numerical modeling of plate sinkage and in situ bevameter shear tests using FEM and the density-dependent NorSand critical state constitutive soil model.

View Paper

Experimental Investigation (2023)

Comprehensive experimental studies on terrain characterization, published in Journal of Terramechanics.

View Paper

Master's Thesis (2022)

Experimental investigation of factors that influence Bevameter terrain response. Master's thesis from University of Pretoria.

View Thesis

Updated ISTVS Standards (2020)

Updated ISTVS standards and terminology for testing in modern off-road mobility research.

View Paper

Fun Projects 🤓

GPU-Accelerated Matrix Multiplication in CUDA vs Standard C++: Theory to Implementation (from Scratch - 1783× Speedup)

CUDA Matrix Multiplication

Low-level CUDA C++ from scratch. Achieved 1783× speedup over naive compiled CPU C++. Covers shared memory tiling, memory coalescing, and kernel optimization.

GitAI: Automated Commit Message Generator

GitAI

A command-line utility that automates the generation of structured, technical, and standardized git commit messages using LLMs. Features diff analysis, conventional commits, and local/cloud inference.

Deep Learning Library from Scratch

Implemented a PyTorch-like deep learning framework entirely from scratch, compiling operations down to machine code. This included custom tensor operations, automatic differentiation, GPU acceleration, and a modular API for building neural networks.

View Git (coming soon)

Fine-Tuned LLMs & RAG Systems

Fine-tuned large language models (DeepSeek-Coder-6.7B, LLaMA 3.1-8B) on domain-specific datasets using QLoRA. Built lightweight, low-rank adapters for efficient deployment on constrained hardware. Developed custom retrieval-augmented generation (RAG) pipelines with LangChain agents and GUI-based assistants for intelligent data and code retrieval.

View Git (coming soon)

Lane Detection White Paper

Authored a white paper on deep learning–based lane detection and lane intrusion detection. Investigated semantic segmentation, convolutional encoder-decoders, and real-time inference constraints for autonomous driving applications.

View Git (coming soon)

Full Self-Driving Car Simulator

Built a reinforcement learning–driven self-driving car model in a virtual simulation environment. Combined stereo vision, accurate vehicle dynamics, and advanced control policies to navigate complex road networks with minimal supervision.

View Git (coming soon)

Multi-Axis Load Cell Design

Designed, built, and calibrated multi-axis load cells using strain gauge techniques. Conducted laboratory experiments for validation, producing reliable force–moment measurements in multi-dimensional loading scenarios.

View Git (coming soon)

FEM–DEM Coupling Algorithms

Developed finite element–discrete element coupling methods for simulating complex vehicle–terrain interaction. Enabled accurate modeling of soil deformation, wheel sinkage, and traction performance.

View Git (coming soon)

MBD Vehicle Suspension Modeling

Modeled vehicle suspension stiffness and damping using a full multibody dynamics (MBD) vehicle model with nonlinear tire models. Analyzed ride comfort, stability, and dynamic response under realistic driving conditions.

View Git (coming soon)

Ground Pressure Experiments

Conducted experimental ground pressure tests on over 100 wheeled and tracked vehicles. Used buried piezometric pressure sensors to capture real-time subsurface stress fields, providing valuable calibration data for terramechanics models.

View Git (coming soon)

Get in touch

Interested in collaborating or learning more about my research?