Technology & Projects

Artificial Intelligence & Machine Learning

TF Diagnose

Machine learning and clinicians working together on the early diagnosis of typhoid fever.

Project 2022
TF Diagnose online platform home page with the headline AI plus clinicians for an improved diagnosis of typhoid fever
My role
Software engineer and ML developer; led the remote development team
Period
2021–2022
Status
Completed
Technologies
PythonDjango RESTAngularTypeScriptMachine learning
Links
Online platform (GitHub)
Data collection tool (GitHub)

The problem

The project set out to support earlier diagnosis of typhoid fever by combining machine learning with clinicians' judgement, in settings where internet connections are slow and unreliable.

My contribution

Worked with a team of biologists at MBOALAB Biotech in Cameroon on an open-source health project. Built a low-bandwidth machine-learning model for early diagnosis, the data collection tool that feeds it, and the project's online platform. Led a remote development team that grew from 4 to more than 9 contributors.

Key features

  • Data collection tool: a form for patient records, a table of raw data and a statistics view.
  • REST API with authorisation and export of records as CSV for model training.
  • Model designed to work where bandwidth is low.
  • Responsive online platform for the project, built from a Figma design.

Results and status

Completed as a fixed-term project between December 2021 and March 2022.

Screens