Me

I'm a London-based machine learning researcher. After stints in biotech and quant finance, I now spend my days working on drone intelligence.

You can get in touch with me by Twitter or email. Links are in the top-right.

Posts

• Post-Mortem Python Plotting: Fixing numerical bugs after they've caused a crash.
• Coronavirus Contrarianism: I can see the logic in the British government’s strategy // Actually, I will go outside after all.
• Quarantine Bread: Bread that's hard to mess up.
• Headlines about coronavirus case counts are worthless: Modelling assumptions are what's important.

Packages

I've published a few small utility packages.

• coolgpus: Custom GPU fan control for headless Linux.
• kvbtests: Heteroskedasticity-and-autocorrelation-robust statistical tests.
• modulepickle: Makes distributed compute feel local by transparently shipping your working directory to the remote machines.
• zonotable: Nails Zotero's paper ingestion onto the side of Notable.

Old Projects

All of my recent major work is NDA'd, but there are a handful of projects from ancient history that are in the open. At this point the code is shameful, but I'm still proud of what they were for the time.

• Segmenting Microarrays with Deep Neural Networks: My first project with neural nets, circa 2014; it's a script for automating one of the labour-intensive parts of reading DNA microarrays.
• Commutes & Rent: My second ever programming project, back in 2014; an interactive visualization of commutes and rent in London.
• A Global Shallow Fluid Model: My first ever programming project, back in 2013; an interactive model of the atmosphere on a rotating planet. Unfortunately the Unity plugin it relies on is now very dead.

This Website

I am a terrible web developer and the blog frameworks I tried were too complex for me to understand. As such, this website is rendered from Markdown into a minimal template by a short Python script.