Samuel Onyambu
Statistician · Researcher · PhD, UCLA
I am a statistician working at the intersection of Bayesian optimization, design of experiments, and applied machine learning. I completed my PhD in Statistics at UCLA in 2024 and currently teach there as an Adjunct Assistant Professor.
I build methods and the software that makes them fast: my research code runs on native C/C++ backends (with OpenMP parallelism) wrapped in R packages, and my applied work spans Python (PyTorch, TensorFlow), R, C/C++, and Fortran.
What I work on
Bayesian optimization & experimental design – KARGO, a trust-region Bayesian optimization algorithm; uniform projection (UniPro), maximum projection, and maximin optimal designs.
Applied machine learning – image-classification projects in PyTorch and TensorFlow (face-mask detection; X-ray fracture screening), alongside work-in-progress explorations in development data science (satellite-imagery poverty estimation, facility mapping, informal cross-border trade) and medical ML (offline reinforcement learning for sepsis with MIMIC-IV).
Scientific software – R packages with native C/C++ backends, including differential evolution optimizers parallelized with OpenMP.
Highlights
- Tuning Differential Evolution Algorithm for Constructing Uniform Projection Designs (with Hongquan Xu), Journal of Statistical Planning and Inference, 2025.
- Invited talks at EcoSta 2025 (Waseda University, Tokyo) and the Loma Linda Workshop 2025; speaking at DAE 2026 (Rutgers University).
- UniPro: design construction via differential evolution with a native C, OpenMP-parallel backend.