As part of a University initiative, I worked as a Backend Developer on a smart receipt budgeting application for Capital One. The purpose of this application was to perform OCR on receipts, categorize items accurately, and provide users with statistical analysis to better manage their budgets, aligning with Capital One's mission of financial empowerment.In this role, I led the backend development within an Agile team. My responsibilities included:Creating a RESTful API in Python using Flask.Developing and maintaining an SQL database.Facilitating interactions between the AI and the database through API calls.Preserving app sessions via session tokens, establishing a many-to-one relationship.Managing user accounts.Handling communications between frontend and backend components via the API.Initially creating the Docker container and subsequently upgrading it to Docker Compose to support multiple containers.Co-maintaining the Docker configuration with a group member to ensure stability and utility.The project achieved a categorization success rate of 85%, utilizing techniques such as fuzzy matching, NLP, and Bayes' theorem. Essentially, we developed an AI that categorizes items with a high success rate, all running on a phone.We managed this project using sprints and conducted team meetings after every two sprints. Every other meeting included the same assigned Capital One employee each time to discuss progress and clarify expectations for the software.At the end of the project, we showcased our application both at the University during an Open-Day and to the members of Capital One in their Nottingham office. Both parties were delighted with the amount of work achieved and the quality of the application.