Alistair Ferag

Senior Machine Learning Engineer at Satalia @Satalia

London, GB

Joined as the first dedicated ML engineer and helped to expand the team to 10 FT data scientists, driving the company to become one of Europe\'s leading end-to-end AI consultancies.• Implemented from research through to production an adaptive probabilistic demand forecasting model to solve a pricing optimisation problem for a global hotelier.• Led the client engagement and development work for a multi-year project with a B2B client where we successfully deployed a product segmentation and customer churn analysis across 17 countries, cutting the identification of churned customers from 90 days to 4.• Developed a footfall predictive model for 92 stores of a nationwide retailer using historical footfall, sales data, seasonality and weather data ultimately leading to an 18% increase in conversion rate.• Designed and led the development for our internal company data lake processing batch and streaming data, utilising a NoSQL database for storage and building a secure API service.• Work extensively with the client to gather and refine their needs and also to understand and mitigate potentially unintended negative consequences of deploying machine learning solutions.• Supervised a total of 15 Master’s students managing the intricate dynamic of delivering results for the client, producing beneficial work for the company and meeting the universities rubric requirements.• Research Areas: Bayesian Machine Learning, Change Point Analysis, RNNs, Causal Modelling• Technical Experience: Python, R, SQL, JavaScript, Docker, Flask, Swagger, Kubeflow, Dash, pandas, Dask, scikit-learn, Keras, Tensorflow, pymc3, ArcMap, Google Cloud (App Engine, Datastore, Cloud Run, Cloud Functions, Compute, Storage, BigQuery, Dataproc).

Lincoln Minster School

A Levels, Physics; Mathematics; History

UCL

Master of Science (MSc), Computer Science

University of Bristol

Bachelor of Science (BSc), Economics

King's College London

Master of Science (MSc), Climate Change