About Miles

When not referring to himself in the third person, Miles likes to play with machine learning, statistical modelling, and data. I’ve worked in domains such as pricing, optimisation, customer analytics, forecasting, natural language processing, asset management, technology architecture and risk management. Most recently my gigs have tended to involve setting up platforms and CI/CD infrastructure using some combination of AWS/GCP, Kubernetes, Kafka, Elasticsearch and Presto, but I did pure statistical modelling and machine learning for several years (before realising that my job would be a whole lot easier if someone would just set up a decent platform to do it on, and that someone may as well be me.)

I like to work across the full stack, and write Python, R, SQL, Scala and Javascript, with experience with technologies and libraries such as Pandas, Numpy, Scikit-learn, Tensorflow, Theano, React.JS, D3.JS, Docker, Kubernetes, Spark, Kafka, Accumulo and many more. I’m particularly interested in Bayesian modelling and neural networks, and combinations of the two are twice as interesting; in the past I’ve worked heavily with probabilistic graphical models, convolutional neural networks, recurrent neural nets (sequence to sequence is a particular passion) and reinforcement learning.

If you’re looking for the latest machine learning paper off Arxiv you’ll be dissapointed, I find that most organisations are still sitting at the point where they need to get good data infrastructure in place, so that’s what a lot of the content on this blog is about (also - snarky takes on the industry, careeer compromising bitterness, and self deprecating humor).