Egor is the Head of AI at Wise, the global technology company building the best way to move money around the world. He first learned mathematics in the Russian tradition, continued his studies with an MSc at ETH Zurich and got a PhD at the University of Maryland. He’s been doing data science before it was a thing, including economic and human development data analysis for nonprofits in the US, the UK, and Ghana, and 10 years as a quant and occasional trader at UBS then Deutsche Bank. Egor has multiple side projects using neural networks for molecular optimization, NLP, and other areas. Following this decade's explosion in AI techniques, Egor was Head of AI at Mosaic Smart Data Ltd for three years, and is now bringing the power of AI to Wise.
Understanding the lifetime value of customers is tricky. They rarely let you know they’ve stopped using your product, so it’s difficult to know when churn happens. At Wise we’re building a neural network model to improve our customer lifetime value and churn probability estimation. At this talk we’ll let you know of ways to estimate the value of the future revenue stream from a customer or group — a vital topic for all retail-facing companies. We’ll start with the naive approaches, explain their challenges, and demonstrate the classic “Buy Till You Die” approach, as implemented in the Python lifetimes package. We’ll then give a brief overview of basic segmentation techniques based on the Recency-Frequency-Monetary value framework. We’ll wrap things up with early results from the neural network model we’re currently developing at Wise, which naturally integrates all of the above. Who’s this presentation for? Data scientists, data analysts, researchers, predictive modelers and anyone interested in machine learning techniques.