Data Science Manager
CurrentLed a team of 3-7 data analysts & scientists through project lifecycle, from clarifying business requirements to delivering effective solutions. - Awarded highest grade in last annual review, given to only top 1% of AXA employees. - Optimised retainer discounts, leading to additional £1M p.a. profit & £4M p.a. revenue. Integrated customer risk and market price predictions. Validated through A/B testing. - Utilised MIB CUE industry database to reduce £200k p.a. application fraud loss. - Prototyping self-serve Gen AI chatbot for data insights. System built using internal metadata, industry-standard definitions, and few-shot examples. - Trained convolutional neural network to detect flat roofed homes using satellite images. - Identified external data sources to incorporate into AXA’s burn cost models to improve COR. Ran XGBoost in PySpark, weight by exposure and offset with current predictions. - Built time series forecast of sales to measure against actual marketing campaign results. - Designed a dash application to visualise the impact of webpages on the customer journey. - Employed fuzzy matching techniques to link quotes and detect abnormal manipulation across applications. Presented it to AXA’s retail COO for approval. - Implemented reporting across cloud storage to ensure data retention compliance. - Organised and hosted an actuary & analytics knowledge sharing event across AXA UK. - Represented AXA at the 2023 Databricks Data & AI summit in San Francisco. - Completed Databricks Lakehouse fundamentals certification.