As Machine Learning Engineer on an industry placement at Capgemini (January 2020 – January 2021), I built a TensorFlow/OpenCV computer vision pipeline for document workflows. The system achieved 92% document field extraction precision, processing 10,000 documents per month in production, and was integrated into Microsoft Teams and Outlook.
I designed a serverless ingestion layer with AWS Textract, Lambda and S3 for unstructured handwritten forms, reducing manual data entry by 80% across client document workflows.
The work also formed Capgemini’s TechChallenge entry: our team, “The Visionaries,” pitched the solution to senior executives and placed Top 3 of 40 teams. The computer vision tool was adopted across Capgemini Netherlands. The project formed the basis of my MSc dissertation at the University of Wolverhampton, for which I received a Distinction Star.