CV

+44 7563 797464
josh@joshredmond.com
https://github.com/Josh-Redmond-UK

Summary¶

I’m a data scientist and software engineer with expertise in using Python, and R to build AI models and analyse data. I enjoy creating web based interactive maps, applications, and visualisations that help users explore these models in JavaScript. In March, I submitted my PhD thesis titled “Participatory Design of Citizen-led Remote Sensing Forensics Tools in Mexico” exploring the use of participatory design to create web GIS software and remote sensing analysis tools.

Education¶

PhD in ‘Environmental Intelligence: Data Science and AI for Sustainable Futures’
University of Exeter
September 2019 - March 2024 (Corrections in Progress)

• Developed obfuscation methods for satellite imagery using classical computer vision and generative AI (diffusion models) methods, implementing scientific papers using TensorFlow in Python.
• Collaboratively designed web GIS tools with stakeholders in Mexico for forensic analysis of satellite imagery in JavaScript (React, Leaflet) and Python (FastAPI, Google Earth Engine, GeoPandas).
• Carried out fieldwork in person in Mexico and online to create the specification for the software, collaboratively designing and iteratively testing with users and stakeholders.
• Developed semantic segmentation methods for detecting human rights abuses in satellite imagery using TensorFlow, presenting this at the IEE GRSS conference.
• Chaired sessions and workshops at the UKEO and GIScience conferences.
• Created new machine learning benchmarking dataset for classification of satellite image classification in Latin America, using AWS and Google Earth Engine to gather and manage data.
• Applied for and received £3000 in funding as a co-investigator for the Public Engagement with Research fund to develop a set of web-based ethical generative AI art tools.
• As part of a team of volunteers, created a dashboard in R Shiny to model the PPE and bed requirements for different COVID pandemic scenarios in North Devon in Spring 2020.

BSc in Politics and International Relations (2:1 Hons)
University of Exeter
September 2014 - June 2017

Experience¶

Data Scientist
Egregious Ltd
June 2024 - Present

Data Science Volunteer
Map Action
April 2024 - Present

Early Career Researcher
UK Health Security Agency
April 2021 - October 2021

• During the COVID pandemic, responded to urgent requests for data analysis and visualisations from ministers and the chief medical officer (CMO) e.g. assessing pandemic trends in problem areas using Python (statsmodels, scipy, matplotlib), SQL, and R.
• Investigated and modelled the impact of the Immensa lab testing failure, visualising and presenting results to stakeholders.
• Led research project into effects of school closures on COVID infection rates with stakeholders across the 4 UK governments, reviewing literature, analysing data, and writing up results which were shared within the department and across the devolved administrations.
• Built a COVID early warning system for rising cases using anomaly detection in Python (numpy, Pandas) methods that was used operationally as part of regular briefings with senior leaders.

Postgraduate Teaching Assistant
University of Exeter
September 2020 - June 2021

• Taught module “learning from data” leading workshops focusing on the theory and application of ML methods including random forests, neural networks, and statistical methods like linear regression.
• Also taught “social network and text analysis” covering basic network and text analysis methods such as community detection and bag of words models.
• Received a teaching assistant award from the department of computer science.

Space Placements in Industry Intern
Earth-i
July 2020 - September 2020

• Created a computer vision based flood risk model in Python (rasterio, GDAL) with TensorFlow which used globally available earth observation and geomorphological data to estimate annual flood risk based on the Environment Agency’s flood risk map.
• Produced and presented a research poster summarising the model and research process.

Research Assistant
University of Exeter
June 2016 - July 2020

• Researched geospatial social networks using R and QGIS, including geocoding data, analysing network structure with statistical models, visualising data, and writing analysis for papers and presentations.
• Presented results at international conferences, contributed writing to peer reviewed publications.

Projects Officer
UCL
Jan 2019 - August 2019

• Designed the program of the Academic Practice and Technology (APT) conference, curating conference sessions according to the submissions we received.
• Organised the APT conference, as part of a team, delivering a successful event approximately 100 attendees from the UK and overseas, including hiring and managing staff and promoting the event.
• Achieved associate fellowship of the higher education academy (AFHEA) in recognition of my learning design work at UCL and previously at Exeter University.

Selected Projects¶

Mapping the UK’s Renewable Potential
Collaborating with Friends of the Earth I led the creation of a data-driven map of potential sites for onshore wind and solar development in the UK using QGIS and the Google Earth Engine Python API using AWS (S3, EC2). This accounted for planning, environmental, and physical constraints. The England portion was published as an interactive map and as part of a policy report by Friends of the Earth, reported on by The Guardian.

Designing Deforestation Monitoring Training for Ex-Combatants in Colombia
Summary of a training course design process training ex-FARC guerrillas in using Global Forest Watch to support ecotourism

Geospatial Web Apps in Python and R
Full Day Workshop I designed and delivered at the International Conference on Geographic Information Science 2023, I later worked with digiLab to develop the Python section as a standalone course.

Shell Offsets Explorer
Hackathon project which used earth observation data to investigate Shell’s carbon offset projects. Visualisations created with Google Earth Engine and Leaflet.js.

LangTree
Large Language Model interpretability tool that visualises output as a tree, built using Python (Hugging Face transformers, FastAPI, inseq) and Javascript (d3.js)