$ whoami

Thomas J. Fletcher

I'm a Machine Learning Scientist at Ground Truth Labs, building deep learning and computer vision pipelines to analyse bone marrow morphology for blood cancer diagnosis and prognosis.

Previously I was a postdoc at Imperial College London developing 3D Gaussian Splatting techniques for fetal MRI, and at King's College London working on inverse problems in cardiac MRI — reconstructing cardiac CINEs from single-heartbeat acquisitions using physics-based deep learning.

I obtained a PhD in Astrophysics at UCL, where I studied the role of the first galaxies in reionisation and the gas content driving star formation in local galaxies. I led an international collaboration that discovered ionising radiation escaping from some of the earliest galaxies in the Universe.

PhD
Astrophysics, UCL
 0
Citations
 0
h-index
 0
Publications
 ...
GitHub Contributions (last 52 weeks)

// experience

Where I've worked

2025 – Present
Machine Learning Scientist
Ground Truth Labs
Deep learning and computer vision for digital pathology. Building cell segmentation and classification pipelines for histopathology, supporting clinical trials in haematology.
2024 – 2025
Postdoctoral Researcher
Imperial College London
Worked on motion-robust and efficient fMRI for infants within the I-X AI initiative, jointly with Trinity College Dublin. Developed Gaussian splatting and machine learning techniques for slice-to-volume reconstruction of highly motion-corrupted MRI acquisitions.
2020 – 2024
Postdoctoral Researcher
King's College London
Machine learning for inverse problems in cardiac MRI. Developed a deep learning framework to rapidly reconstruct CINEs acquired in a single heartbeat, using networks for estimating non-rigid cardiac motion and physics-based MRI reconstruction. Built invertible neural networks for quantitative cardiac magnetic resonance fingerprinting.
2016 – 2020
PhD Astrophysics
University College London
Thesis: The Ionising Output and Gas Content of Galaxies. Led an international collaboration using ground and space-based telescopes to discover leaking UV radiation from early galaxies. Used Bayesian statistics to measure the cosmic abundance of molecular hydrogen in the local Universe.
2015 – 2016
MSc Astrophysics
University College London
Dissertation: The Cosmic Abundance of Molecular Gas.
2015
Research Scientist
National Physical Laboratory (NPL)
Part of the Thermal Imaging group. Delivered accurate and traceable 3D thermal imaging capabilities, including rendering and ray-tracing, for satellites undergoing testing at the European Space Agency's ESTEC facility. Worked on a clinical trial comparing a novel thermal imaging device against non-contact infrared thermometers for diabetic patients.

// publications

Research output

→ View all on Google Scholar

// skills

Tech stack

ML / AI
PyTorch TensorFlow scikit-learn segmentation-models-pytorch MLflow Pydantic Computer Vision
Languages
Python MATLAB TypeScript JavaScript SQL R Bash LaTeX
Scientific Computing
NumPy SciPy Pandas Matplotlib Seaborn plotnine scikit-image Shapely OpenCV Jupyter
Tools & Infrastructure
Git Docker GCP Linux CI/CD uv Ruff Next.js Tailwind CSS React
Domains
Histopathology Digital Pathology Cell Segmentation Clinical Trials Inverse Problems Gaussian Splatting MRI Medical Imaging Astrophysics Cosmology Metrology
AI-Assisted Dev
Claude Code Cursor GitHub Copilot Claude Skills

// contact

Get in touch

~/contact
$ echo $GITHUB → github.com/tomjf
$ echo $LINKEDIN → linkedin.com/in/tomjfletcher
$ echo $SCHOLAR → Google Scholar
$ echo $LOCATION → London, UK