About Me

Work Experience

• (2021-present) Data Scientist at Arca Blanca. This being a role at a consulting firm my work has been extremely varied across a number of different sectors
  • Customer Lifetime Value project with FTSE 100 software firm
    • Built live ETL pipelines
    • Created survival regression model of customer lifetime
  • Healthcare data analytics system Project
• (2019-2020) Data Scientist/Quant at Smarkets
  • I worked in the quantitative department of electronic market making at Smarkets, researching and developing trading algorithms on our exchange.
  • I implemented a bot for pricing and trading multiples (products of betting contracts) as a flask app, that ingested live data from a number of betting price sources including other strategies on the exchange.
  • I worked in a cross-departmental team to create a data classes capable of performing strategy-wide bottom up analysis on each execution.
  • I contributed to company wide knowledge repositories with notebooks detailing my pricing research.
  • I contributed to the creation and maintenance of ETL pipelines
  • I wrote jobs to train the machine learning algorithms required to populate our football team strengths pricing model using scikit-learn and tensorflow, including adapting cutting edge research papers to improve their efficacy and accuracy.
  • I was on a tech ops rota, being responsible for dealing with live issues in time critical environments, and as such managing our company kubernetes cluster, making codebase changes and deploying changes with crane and k8s.
• (2018-2019) Blenheim Chalcot Data Science Graduate Program/Instrumental
  • I worked as a data scientist and machine learning engineer for the startup Instrumental under the Blenheim Chalcot venture builder portfolio Umbrella.
  • I implemented Knowledge Graph techniques to predict artist similarity, clustering methods (such as the Lance Williams algorithm) to measure genre similarity
  • I developed a model which predicted which spotify playlists drove ticket sales the most using elastic net regression, and predicted which audio features drove streaming popularity in songs
  • I developed a model using one dimensional convolutional neural networks to read the copyright text of an album to determine its record label, each of which are novel applications of their respective techniques.
  • Each of the above were put into production through both in house linux servers (using software like Cron) and AWS lambda/EC2, for which I took on the work of engineering myself. Capital One Business Analysis summer internship (2017)

Education

Undergraduate study:

(2015-2018) Girton College, Cambridge, MA Mathematics (automatic upgrade).

Specialised in:

• Statistical modelling
• Differential geometry
• Topology.

Postgraduate study:

(2018-2020) Birkbeck College, University of London, MSc Mathematics.

Specialised in:

• Financial Mathematics
• Group Theory
• Graph Theory
• Topology

Dissertation on using Topological Data Analysis for model selection in dimensionality reduction and manifold learning.

Go back home.