Garfield is an Legal AI startup aimed at increasing access to justice by enabling unrepresented parties and smaller businesses to pursue small track claims. I have been developing the backend of this application within a small team. We're hoping to be able to share more soon!
Data Scientist
ElectronRx
2019 - 21
I previously worked as a data scientist at a Cambridge based biotech start-up. This role entailed developing predictive algorithms from signals captured by smartphones, as well as in-house developed devices. The primary focus was on developing an algorithm to infer blood pressure from a ppg (photoplethysmography) signal, taken by a smartphone camera. This involved harnessing approaches proposed in recent literature to develop a sufficiently accurate machine learning model and effective signal processing pipeline. The work was then integrated with the wider app platform (FoneDx), which aimed to allow any modern smartphone to become a medical grade device. In this position I also took significant responsibility in maintaining and developing the backend for the platform, as well as in the data engineering tasks of designing and implementing a data pipeline which used cloud computing services such as AWS S3 and AWS Lambda, allowing us to process terabytes of biomedical data.
Over a two-month period I worked to produce an account application fraud detection model. The model was trained to score applications as either fraudulent or safe. The model had many components, including features generated from user sequence analysis, computed using Markov chains, an entity resolution component and a more general network analysis components. The features were then passed to an XGBoost classifier which achieved an AUC of over 0.9.
Technical Lead
CleanPlate
2021 - Present
In my spare time I have been working with friends on a non-profit pushing for carbon labelling on food packaging. To push for this we are developing a mobile app which enables food influencers to easily label their recipes and share insights into the carbon impacts of the ingredients with their audiences. By enabling these individuals we believe we can reach a large audience and help inform their dietary choices. The app is currently in a final user testing phase before release, we have also successfully been awarded several grants, which will enable the promotion of the app.