Education
Imperial College London
PhD in AI and Machine Learning
2025 - Present
Thesis Project:
Graph Neural Networks for scoring single-cell genetic health
Research and development of bespoke graph-based representation learning methods and probabilistic graph ensemble generation from partially observed data.
Warwick University
MSc in Statistics (Distinction)
2022 - 2023
Thesis Project:
AI-informed Approximate Bayesian Computation in decision field theory
Developed deep learning based automatic summary statistic selection method for approximate Bayesian computation applied to decision field theoretic diffusion modelling of human decision making data.
The Open University
BSc in Mathematics (First Class)
2017 - 2021
Research Project:
Ergodic theory and applications of probabilistic frameworks to chaotic systems
Study in asymptotic behaviour of simple chaotic systems, benefits of applying probability theory to chaotic spaces, and historical development of chaotic psuedo random number generator algorithms.
Current Projects
Imperial College London
2025 - Present
Python Package Development (pySPoC)
Developed extension for DARTS neural architecture search algorithm for convolutional neural networks for joint search over network weights and quantity of channels. Method was used for image-based cancel-cell detection.
Imperial College London
2025 - Present
Automated Feature Extraction for Approximate Manifold Detection
Developing automated feature extraction methods for manifold approximation and dataset classification for empirical study of the Manifold Hypothesis.
The Charité - University Hospital Berlin
2025 - Present
EEG Signal Extraction
Investigating improved power spectrum recovery and viability of relaxed experimental setups afforded by newly developed signal extraction method for haptic EEG data.
Past Projects
Warwick University
2023
Differentiable Neural Architecture Search
Developed extension for DARTS neural architecture search algorithm for convolutional neural networks for joint search over network weights and quantity of channels. Method was used for image-based cancel-cell detection.
Warwick University
2023
Options Pricing Dynamics Study
Exploratory study of path-dependent stock options risk-neutral pricing dynamics under varying market conditions and computational limitations inherent in Monte Carlo and finite difference method pricing.
Industry Experience
Commify
Data Scientist
2023 - 2025
Nottingham, UK
Customer Churn Prediction
Reduced year-on-year customer churn by 4% by developing PySpark churn prediction pipeline, utilizing STL decomposition for seasonal customers, Kalman filtering for interim forecast outputs, and XGBoost classification ensemble over varying time horizons.
Embedded Image Filtering in RCS Messaging
Improved illicit image filtering in RCS traffic by 60% utilizing convolutional neural network for binary image classification with production model deployed as PyTorch inference pipeline.
Upsell Improvements via Collaborative Filtering
Increased year-on-year upsell by 9% by developing automated upsell suggestion pipeline based on collaborative filtering applied to customer and product data with automated monthly opportunity reporting in Databricks.
Data Transformation Platform Development
Reduced data engineering team workload by 10% by developing configurable data transformation platform back-end using PySpark, SqlAlchemy, and Spark-SQL.
IRIS Software Group
Data Scientist
2019 - 2022
Slough, UK
Customer Whitespace Analysis
Increased cross-sell rates by 12% YoY through customer white-space detection, implementing agglomaterative clustering on kernel PCA latent customer data in Python, with configurable cluster cut-offs.
Product Bundling via Conditional Random Fields
Developed product suggestion pipeline for sales and commercial team through PyTorch inference pipeline, utilizing conditional random field model over product nodes, and automated reporting of graph cliques as product bundles in Spark.
IRIS Software Group
Data Engineer
2017 - 2019
Slough, UK
Merger and Acquisition Systems Onboarding
Reduced on-boarding time for new M&A data feeds by 70% (averaged across 7 M&As) by designing and developing .NET C# application providing configurable, adaptive, and automated data ingestion pipelines.
CI/CD Automation and Pipeline Development
Reduced release overhead by 20% and reduced frequency of release incidents by developing Azure DevOps CI/CD release pipelines with Windows Powershell, Bash, and YAML.
MasterCard
Python Developer
2015 - 2017
London, UK
Hedging Tools Development
Improved hedging decision making by developing internal analytics tools for currency exposure tracking and hedging analysis with Excel and Oracle Database integration.
Document Scraping and Oracle Database Integration
Reduced regulations team workload by 15% by developing document scraping application with Excel VBA interface.
Skills & Proficiencies
Programming Languages
Python (Expert)
SQL (Expert)
R (Intermediate)
C# (Intermediate)
Java (Basic)
C++ (Basic)
Technologies
-
Microsoft Azure
-
AWS
-
GCP
-
Apache Spark
-
Docker
-
.NET
-
Slurm
-
Git
Packages
-
PyTorch
-
XGBoost
-
Numpy/Scipy
-
Pandas/Polars
-
MLFlow
-
Cython
-
Torch for R
-
Tidyverse
Research Focus
-
Probabilistic Deep Learning
-
Computer Vision
-
Neural Architecture Search
-
Representation Learning
-
Bayesian Statistics
-
Optimization