Projects
8 case studies
AHT reduced by 2 mins avg Call Analysis Deep Dive
Average handling times at British Gas dropped by two minutes after targeted agent retraining, delivering substantial OPEX savings across the contact centre.
£2M annual savings Engineer Job Trial
An end-to-end ML model in production on Azure ML and Databricks, reducing allocated engineer job times by 15% and saving £2 million annually.
£10k profit generated Fuel-level & Trade Optimisation
A collaborative R forecasting package built with Centrica and NEAS Energy's ATS team in Denmark, generating £10k in additional profit.
£200k annual savings Wrap Notes Automation
AWS and Salesforce integration applied a summarisation model to webchat transcripts, replacing manual wrap notes entirely and saving £200k per year.
Algorithmic Trading Bot
An algorithmic trading bot placing optimal virtual trades, which later served as direct inspiration for pursuing the Oxford Algorithmic Trading Programme.
Repeat visits reduced Predictive Maintenance
An R Shiny app guiding field engineers on optimal part selection, presented at senior Services and Logistics meetings and credited with reducing repeat engineer visits.
HH Short-term Forecasting
Half-hourly Smart Meter data used to predict UK power market settlement values, supporting real-time grid optimisation and balancing.
Twitter Dashboard
A real-time dashboard tracking Hive brand mentions, sentiment, and engagement signals to surface intelligence for product and communications teams.