About
I’ve been leading the ML and Data Science team at Quine since late 2020. Quine is a small startup with the aim of quantifying developer competence and reputation. The goal is twofold: to help developers monetise their experience via job, contract and issue matching, and to make the open source ecosystem more sustainable. Prior to this, I completed a PhD in ML (and some post-doc work) at the Alan Turing Institute (supervised by Chris Williams) after spending 5 years as a data scientist in retail and credit risk.
I have two primary professional interests: the insight and acuity that analysis and algorithmic thinking (maths, computer science etc.) provide to technical problems, and the creation of novel and delightful products. For me, each interest spurs on the other. I believe there is much more that ML and statistics can unlock creatively in the world, although I would be more cautious in my claims than the loudest voices in the media. Some career highlights to date:
- Hierarchical time series decomposition with partially-known time-based effects for supply chain data (Tesco - proprietary).
- On-the-fly personalisation for drug response monitoring (PhD work — see research).
- Interpolatable walking style for mocap video generation (PhD work — see research).
- Creating a taxonomy of software engineering, and a bootstrapped model to automatically tag repos on GitHub for organisation and discovery (at the time of writing we have auto-tagged more than 100m GitHub repos).
- Recommendations for repositories and issues for developers on GitHub (blog post coming soon).
On the engineering side, I'm experienced in OLAP databases and analytics, data engineering and system design, backend, cloud infra, scraping, and visualization.