NatWest Group (Client)

I worked within Retail Analytics and Decisioning team, where, amongst many other things, I

• Developed three propensity to upgrade models targeting lower-paid account customers for upgrades; utilized decile-based response rates and rigorous risk governance; generated £100K+ in benefits; productionised the model for delivering monthly outputs to stakeholders.

• Implemented a causal forecasting model integrating customer feedback, transactional data, and external factors; proactively addressed potential mis-sellings, saving £0.5 million; showcased at NatWest’s mini Data Science and Engineering Conference 2023.

• Developed a strategic data resource using Teradata SQL, Snowflake, and Hive; identified and mitigated criminal behaviour in personal accounts used for business transactions.

• Developed uplift models to identify customers most likely to upgrade from non-fee to higher-paid accounts given previous customer campaigns; resulted in a 30% increase in the packaged accounts customer base in 2023.

• Employee of the month: December 2022, January 2022, and August 2023

Research and Development (Internal)

• Led an intern project on generating synthetic data for financial transactions using conditional and generalised adversarial networks, and autoencoders.

• Mentored two non-data colleagues and two junior data scientists in their Data Science and Python journey.

• Developed a Hybrid LSTM-Random Forest model for forecasting electricity consumption at UK EV charging depots; optimised new charging point locations with a location-allocation model considering distance constraints and existing infrastructure.

• Received 10+ Gold, Silver, and Bronze Employee Recognition Star Awards for contributions to projects, mentoring, and community engagement.

Tech stack : Python, AWS Sagemaker, PySpark, Git, Snowflake, Bit bucket, Teradata, Databricks, inner source

• Worked with Dr Gonçalo dos Reis at the Battery Lab, University of Edinburgh to design a python package to clean several battery datasets, to detect the rainbow, and full cycling for Lithium ion batteries, using various algorithms and unsupervised machine learning methods.
• Managed collection of further data related to Lithium ion batteries for research purposes.

Tech stack: Python