I’m interested in machine learning systems, parallel computing, and performance engineering. This profile is a curated collection of projects where I explore how software behaves at scale through measurement, optimisation, and reproducible experiments.
| Project | What it explores |
|---|---|
| Scalable Vision Transformer training | PyTorch input-pipeline tuning, mixed precision, torch.compile, gradient accumulation, and multi-GPU DDP |
| MPI predator–prey simulation | Actor-style message passing, distributed simulation, Slurm experiments, scaling, and validation |
| Molecular-dynamics performance | Portable and architecture-aware optimisation of an O(N²) C kernel |
| Compiler optimisation study | The performance and numerical effects of C compiler optimisation modes |
| IMDb analysis in Scala | Typed data processing and tested analysis over IMDb datasets |
| Chain-of-Table reproduction | A clearly attributed reproduction of the archived Google Research project |
Each featured repository includes setup instructions, its current validation status, and caveats around the recorded results. Some projects began as coursework or research experiments, so please check the repository’s licence and any data terms before reuse. If no licence is present, ask before reusing the code.
Questions, feedback, and collaboration ideas are welcome. Open an issue in the relevant project or reach me through GitHub.