Matthew Richard

Senior Data Scientist, Kinduct

The statistics program at SFU provided me with a solid foundation and framework from which my career in statistics could grow. The diverse range of statistical fields provided me with enough knowledge to confidently seek out employment as a budding statistician. The professors in the SFU statistics program were enthusiastic, knowledgeable and always had time for a 1-on-1 discussion regarding statistical theory.

Since the completion of my degree at SFU, I’ve worked at many different companies under various job titles. While the job titles varied considerably (data analyst, data miner, data scientist), all the positions were similarly grounded in statistical theory and training.

My career started as a volunteer at Vancouver’s Women and Children’s Hospital working as a statistician. I performed analyses for non-statistical researchers who had specific hypotheses about their data.

I was later employed as a data analyst at a large mental health services company in Portland, Maine. It was here that I immersed myself in R, exploratory analysis, and developing efficient methods for dealing with unstructured data.

After 2 years in Portland, I took a position as a data scientist at a medical data company based in Halifax, NS.  This position focused heavily on large, sparse binary datasets with hundreds of variables. This position required strong knowledge of algorithmic techniques for dealing with the curse of dimensionality.  I became heavily influenced by Bayesian statistics and machine learning techniques. By the end of this position, my tool belt included clustering techniques, neural networks, decision forests, multivariate reduction techniques, subgroup analysis, and statistical simulations. This position had a very academic-like setting, and provided the opportunity for many publications, posters, and international travel.

In 2015, I left the medical field and joined Kinduct, a sports data company in Halifax, NS. This position requires the use of machine learning techniques to derive meaningful metrics from an ever-increasing amount of data.