Matthew Bracher-Smith

Background

About Matthew Bracher-Smith

I am a Bioinformatician and Data Scientist bridging the gap between complex biology and scalable machine learning. With a PhD in machine learning (ML) for genetic prediction, I specialise in transforming massive, noisy healthcare datasets into robust predictive models.Currently, I build predictive models for Alzheimer's disease using large-scale genomic data and Electronic Health Records (EHRs). My work relies heavily on learning from complex and noisy biological data, building reproducible ML pipelines, causal inference in healthcare, and robust survival analysis. I am deeply passionate about low risk-of-bias ML model development and reporting.Beyond modelling, I prioritise clean, scalable code and effective communication. I have spent 8 years teaching R for data science (specifically the Tidyverse) to MSc students, translating complex statistical concepts into practical workflows. I am currently focused on optimising complex SQL queries and expanding my knowledge in C.Technical Stack & Expertise:Machine Learning: Scikit-learn, PyTorch, XGBoost, Dask-ML, SVMs, Penalised Regression, SHAPProgramming & Data Engineering: Python (General SciPy stack: NumPy, Pandas, Dask, Polars, matplotlib etc.), R (Tidyverse, and modelling libraries), SQL.Domain Knowledge: Neurological and psychiatric diseases and disorders, GWAS, Polygenic Risk Scores, Survival Analysis, Causal Inference, EHR Extraction.