Revolut
From 2021 to 2024, I led product strategy, operations and data teams at Revolut. My specific focus was leading teams responsible for user lifecycle, engagement and retention. During my time, I was interviewed as part of the Revolut Insider podcast to discuss Revolut's strategy and how we built product for our +45M customers worldwide.
Side Projects
May 2021
Machine Learning for Monitoring Forest Restoration
For my undergraduate dissertation, I worked in collaboration with researchers from the environmental non-profit organisation Conservation International to develop a machine learning model to monitor a forest restoration project in northern Fiji using satellite imagery.
My proposed machine learning model achieved 96.4% classification accuracy in identifying the presence of trees in satellite imagery and was shown to be successful to help quantitatively assess changes in forest cover between distinct time periods. My dissertation was shortlisted for outstanding achievement by The University of Edinburgh.
April 2021
Predicting US Droughts with Machine Learning
In my final year of undergraduate studies, I was awarded 1st Prize IBM Award in Machine Learning for an paper that I co-authored. Our paper used deep learning AI models to predict and forecast drought conditions in the United States using meteorological data. Our results demonstrated that deep learning models - LSTMs and Transformers - can be successfully used to forecast drought conditions which are crucial for early-warning drought management systems in the US.
Our paper was accepted and presented at the 2021 International Conference for Machine Learning (ICML 2021) workshop for Climate Change AI. We also open-sourced our dataset, making our unique dataset available on Kaggle.
April 2021
CovidMaps: Mapping Covid-19 Data in Scotland
In May 2020, at the height of the first wave of Covid-19 in the UK, I co-developed the first data visualisation website in Scotland to help compare Covid-19 data across local authorities and health boards. The aim of the project was to provide data in a clear, simple and digestible format, allowing people to easily compare the situation across local regions in Scotland.
The website gained wide coverage, attaining thousands of unique viewers and endorsement from The University of Edinburgh.
April 2021
BetterReads: Building a Book Review Search Engine
As part of my Text Technologies for Data Science coursework, I worked in a team to develop a book review search engine called BetterReads. The system searched a database of 69 million book reviews to help book lovers find new titles based on their review search queries. The technical system combined sentiment analysis machine learning models, with a BERT-based Transformer model, to find the perfect book match depending on the users search query.
We were awarded 1st Prize for our project, winning in a class over 100+ students.
March 2021
Scratch from Scratch: Online Coding Club for Kids
I led a team of volunteers to create a 5-week online coding club for kids aged 9-11 called Scratch from Scratch. The lessons helped introduce 100+ kids across Scotland to programming for the first time.
I was shortlisted for The University of Edinburgh award for Community Impact Award and the project received press coverage including being featured on British Science Week 2021.
Dec 2020
TryAI: Machine Learning for High School Students
I developed an AI education website for high school students aimed at ages 14-16. The goal was to create fun content that helps introduce younger learners to important concepts in data science, machine learning and AI ethics.
In developing the project, I worked with the former MIT Media Lab student Blakeley Payne, whose work on building a data ethics & AI education curriculum for middle school students was been a leading inspiration in developing the TryAI project.
Apr 2020
RemindDoor: Building a Smartlock Prototype
For the Edinburgh University System Design project I managed a team of 10 students and launched an assistive hardware product called RemindDoor. The smart lock prototype was developed with the vision of helping a diverse range users - from health carers, to Airbnb hosts. At the end of the project we pitched our prototype to investors and representatives from companies including Dyson, Amazon and Github.
Our team was awarded 1st Prize in a competition with 200+ students and we won the Informatics Business Development Award for business leadership.
Dec 2019
Developing an Autonomous Drone Flight Path Game
As part of my Informatics Large Practical coursework, I was given the task to develop an autonomous drone flight system in a location-based strategy game.
This project was my first time programming in Java and my software package received a top-grade, attaining 90% for my code design & implementation and 84% for my final overall report mark.
Problems to be solved
There are many critical problems to be solved that can be addressed by better technology. Below are some projects I think about often. If you are working on any of the following ideas, please reach out to me and I'd be happy to support in any capacity.
Building a central platform for the the world's knowledge
Knowledge is fragmented. We are data rich, insight poor, and lack a unifying platform to connect the world's information to make leading knowledge more accessible to everyone.
Wealth management 2.0
The wealth management industry is ripe for transformation. I believe that human-led, financial advice will remain as important in 20 years as it is today. There is a case that advice becomes even more important, with increased diversification and democratisation of financial assets, increasing the number of individuals that require financial advice over the coming decades. The major problem that needs to be solved is to build the world's modern, technological software and data platform for wealth management businesses, enabling best-in-class financial advisory and intelligence to support the management of over £120 trillion in global assets across the world.
Forms of New Government
If we were to establish a new society today, what optimal government structure would we choose? I'm confident our current archaic, 18th century-derived system of governance wouldn't be it. Is it possible to use cutting-edge technology to run leaner governments with greater decentralisation?
I'll aim to add to this list through time. List to be continued...
Other good resources of problems to solve, include lists from Tyler Cowen and Gwern.