Experienced developer with a specialisation working within scientific and regulated environments where I research, design and implement software and data science solutions. I have a first class BSc in Physics and an MSc with distinction in Data Science. My background in Physics complements my Data Science skills and allows me to excel at working within scientific domains. I aim to understand the physical systems involved and capture subject matter expertise into analytical solutions. I’m therefore knowledgeable with, Synthetic Aperture Radar (SAR) Liquid Gas Equipment (LGE), and mass spectrometry.
Hello! I'm a Data Scientist, experienced software developer, rock and synth-wave enthusiast, tennis player and a long time suffering Ferrari F1 fan; currently based in Edinburgh.
I've recently completed an MSc in Data Science at the University of Edinburgh. For my dissertation entitled: "Segmentation of Windthrow in High Resolution Capella SAR Images Using Fully Convolutional Networks" I worked part time as a data scientist at Earth Bloxand closely with Capella Space, achieving a grade of 87%. I'm currently in the process of getting this work published.
My ultimate ambition is to pursue a career where I can utilise my skills to to tackle some of our biggest environmental challenges. I'm currently pursuing this at Sylverawhere I combine deep learning and satellite data to bring transparency to the carbon markets. Before that I worked on the ecoSMRTsystem to minimise CO2 emissions and help the shipping industry meet it's net zero commitments.
Previously I worked for analytical instrumentation companies where I developed instrument control software and data science solutions. I am most proud of improving the self- diagnostic abilities of the instruments and automating previously complex setup procedures which made these instruments more accessible to less technical users! The systems I have contributed towards have been recognised as part of the solution for the Pfizer – BioNTech COVID-19 vaccine!
2021 - 2022
Developed a thorough understanding of machine learning/statistical methodologies and how to apply these to scientific data. For my dissertation, Segmentation of Windthrow in High Resolution Capella SAR Images Using Fully Convolutional Networks, I developed tools to automate the processing of SAR imagery into a format suitable for deep learning and developing novel segmentation algorithms. This work was supervised by Prof Iain Woodhouse and I worked closely with Capella Space. Courses Included:
Achieved an MSc with Distinction
Dissertation Mark 87%
2014 - 2017
Gained an understanding of many physical systems and excellent mathematical abilities in areas covering calculus, linear algebra, mathematical reasoning, probability and statistics.
Achieved a First Class degree
Developed a Least-Squares Fitting Routine Python application currently in use by the University for data analysis and received formal recognition from the Laboratories Committee for the ‘outstanding’ work done and its usefulness for studies in Undergraduate Laboratory work
2013 - 2014
Progression from foundation year required an average year grade of 80% which was achieved by averaging 93%; amongst the highest in my cohort.
Recipient of the Gillett Foundation Studies Scholarship, this was awarded in recognition of my achievements on foundation year
April 2023 - present
Machine learning engineer focussed on combining deep learning and remote sensing to bring transparency to the carbon marketand incentivize real investment in climate action.
Casually doing my best to save the planet 🌳
Researching and developing novel deep learning algorithms to be applied to satellite data
Engineering ML pipelines according to software engineering best standard to ensure they are robust, scalable and extensible.
Liaising with stakeholders to ensure our ML solutions are actually brining value to customers
October 2022 - March 2023
Leading the development of data processing pipelines and methodologies to analyse real time data streamed remotely from Liquid Gas Equipment (LGE) instrumentation.
Leading the deployment of the infrastructure and practices required to productionise data science solutions into commercial products. This is a new endeavour for Babcock and I am liaising with business wide stakeholders to ensure solutions work across the business and remain robust
Developing machine learning and artificial intelligence solutions to minimise operational downtime and ensure equipment is running efficiently as possible; aiding the shipping industry in meeting its net zero commitments
Developing digital twins of instrumentation to aid process (chemical) engineers understand how systems are operating in the field
March 2022 - August 2022
Data scientist at a fast paced startup aiming to make earth observation accessible to everyone, regardless of expertise. I support the development of new earth observation workflows by developing data science pipelines and prototyping new algorithms.
Collated historical windthrow events from the Copernicus Emergency Management Service to facilitate development of new windthrow algorithms in Sentinel-1 data.
Developed automated workflows to scrape, encode and create training data from available earth engine functions and image collections IDs. Developed algorithms to recommend useful earth engine collections to users based on previous workflow usages.
Dec 2020 - Aug 2021
Responsible for feasibility studies and driving the adoption of new technologies/practices. Worked with C#, Azure, Python, Specflow, Gherkin, Appium and more!
Drove the adoption of modern practices and held workshops, demonstrations and presentations
Developed a test framework with the capability to automate at least 50% of the existing manual testing
Incorporated Azure into the existing infrastructure to improve CI/CD
Nov 2017 - Dec 2020
Responsible for the design and implementation of Mass Spectrometer control software. Working within a microservice architecture with C++, C#, Python, Lua, Docker, Angular2+ and more!
Created deep simulations of real time data and instrument behaviour
Developed self-diagnosing and automated instrument setup procedures
Pioneered the use of python for developing new data analysis techniques
Investigated how machine learning could be utilised for predictive maintenance
Jun 2016 - Sep 2016
Involved in the development process and testing of a new mass spectrometer.
Carried out component and sub-system testing on a prototype mass spectrometer
Gained knowledge in maintaining mass spectrometry systems, specifically with GC-MS systems
Suggested design improvements in relation to potential faults and liaised with mechanical engineers in Singapore about these issues
© 2021 Abie Marshall. Built by me! Using Gatsby + React + Chakra UI
Icons made by Freepikfrom www.flaticon.com
Backgrounds customised at SVG Backgrounds