About
I'm Ken Reid — a Senior Data Scientist at Rocket Mortgage with a Ph.D. in Artificial Intelligence from the University of Stirling. My research spans evolutionary computation, optimization, and machine learning, with publications at top venues including GECCO and IEEE SSCI. I design and build data-driven solutions across Python, Java, JavaScript, and more — from automated testing frameworks to large-scale optimization systems. I'm currently pursuing a master's degree in Organizational Leadership, pairing the technical side of my work with the people side of leading teams.
Research Interests
Keywords
Publications
-
An interdisciplinary outlook on large language models for scientific research
-
Heuristic hyperparameter optimization of deep learning models for genomic prediction
-
Optimizing LLVM Pass Sequences with Shackleton: A Linear Genetic Programming Framework
-
A hybrid metaheuristic approach to a real world employee scheduling problem
-
BatchPlanet: Batch access and processing of PlanetScope imagery for spatiotemporal analysis in R
-
Temporal Coupling Between Linguistic Arousal and Heart Rate Variability During Emotional Sharing
-
Genetic improvement in the Shackleton framework for optimizing LLVM pass sequences
-
Benchmarking of the Shackleton Genetic Programming Framework for LLVM Compiler Optimization
-
The Factory Must Grow: Automation in Factorio
-
Metaheuristics For Solving Real World Employee Rostering and Shift Scheduling Problems
-
Shift Scheduling and Employee Rostering: An Evolutionary Ruin & Stochastic Recreate Solution
-
Variable Neighbourhood Search: A Case Study for a Highly-Constrained Workforce Scheduling Problem
Doctoral Thesis
Metaheuristics for Solving Real World Employee Rostering and Shift Scheduling Problems
Real-world employee rostering and shift scheduling, solved with state-of-the-art metaheuristics: Variable Neighbourhood Search, Evolutionary Ruin & Stochastic Recreate, and hybrid matheuristics combining metaheuristics with Integer Programming, evaluated against real-world data provided by BT.
Read the full thesis (PDF, 6.7 MB) →Featured Talk
"The Factory Must Grow: Automation in Factorio"
Portfolio
Projects and interactive tools
Things you can run
Hosted free on Streamlit, so an app nobody has used for a while opens on a wake-up screen: press its button and give it half a minute.
Steam Stats Visualized
Paste your Steam profile and get your gaming life roasted, with stats and shareable cards.
Open the appSource
GoodReads Reading Stats
Upload your Goodreads export for your reading personality, genres, ratings and a quotes tool.
Open the appSource
Debt Payoff Simulator
Avalanche against Snowball, with charts, balance transfers and bi-weekly payments.
Open the appSource
Letterboxd Roasted
Upload your Letterboxd export and get roasted: film personality, contrarian score and more.
Open the appSource
Research, Live
My published research, rebuilt to run in your browser, with the parameters in your hands.
Algorithms, Live
Algorithms running live, with sliders to cause their famous failures yourself.
Code and notebooks
EasyViewerPDF
An Electron app to split PDFs and compare pages side by side.
FactorioBeltProblemGECCO
Evolving Factorio belt balancers: the code behind the GECCO '21 paper.
VAE for Molecule Discovery
A variational autoencoder that generates new molecular structures.
CNN X-Ray Classifier
Pneumonia detection from chest X-rays, as a deep learning tutorial.
NLP Text Analysis Toolkit
Topic modelling, sentiment and classification for any text.
Simulated Annealing
A Colab demo that anneals its way to a target string, step by step.
Introductory Data Science
Colab tutorials from regression and clustering to CNNs and RNNs.
Global Temperature Analysis
R, SQL Server and Tableau on long-term global temperature trends.
Blog
Technical write-ups and project deep-dives.
Let's Collaborate
The fastest route is email: ken@kenreid.co.uk, or the contact page. New posts on data science land in the free newsletter. Profiles, for the diligent:
GitHub
Google Scholar
LinkedIn
ORCID iD
ResearchGate