About

I am an Accredited Statistician, Lead Data Scientist, and a final year PhD student in statistical machine learning in the Dynamics and Neural Systems Group at The University of Sydney under the supervision of Dr. Ben Fulcher. I am principally interested in methods and applications of time-series features—properties which can reduce a time series down to a single summary statistic (or set of summary statistics) which can then be used for machine learning applications—such as classification problems. My research has largely focused on systematically applying and comparing diverse time-series feature sets on applied problems from neuroscience to process engineering to human wearables, and, more recently, developing new efficient search algorithms for finding informative time-series features for any given input data.

A substantial portion of my work (and probably my deepest academic passion!) has involved building open-source software. So far, throughout my PhD I have developed timegpy (genetic programming for finding informative time-average features for time-series classification), theft (Tools for Handling Extraction of Features from Time series), theftdlc (theft ‘downloadable content’), Rcatch22 (calculation of 22 CAnonical Time-series CHaracteristics), correctR (corrected test statistics for comparing machine learning models on correlated samples), normaliseR (rescaling methods for numerical vectors and time-series features), PosteriorPlots.jl (graphical tools for Bayesian inference and posterior predictive checks), and QuasiGLM.jl (adjustments for Poisson and Binomial Generalised Linear Models to their quasi equivalents for dispersed data).

Outside of my direct research, I am also working on GAM.jl—an ambitious attempt to build generalised additive models in native Julia using as similar of an interface as possible to R’s incredible mgcv package. I am also an avid builder of deep neural networks in PyTorch, where my expertise ranges from transformers for large language model architectures (yes, I have developed my own GPTs…) to long short-term memory networks for time-series data.

Professional experience

I am an Accredited Statistician and the Lead Data Scientist at the consulting firm Nous Group where I have worked for 7 years. My role is quite diverse, but I most often find myself working on large and complex evaluations which require data linkage, causal inference, and rigorous statistical modelling, as well as projects requiring forecasting or machine learning methods. I also build machine learning products trained on deep neural networks for both internal and client use, develop internal software packages for Nous, and help maintain its large SQL data warehouse.

Outside of work

I can usually be found writing and producing music for my instrumental symphonic metal project The Archon Rift, running, being a dungeon master for D&D campaigns, or playing far too many RPGs like Baldur’s Gate 3.