My role on this placement was to develop a platform that streamlines the process of finding suitable support agents for machine learning tasks in a secure way that also maintains data privacy. The mechanism behind the platform was inspired by Professor Pinson’s recent work, and the placement was treated like an early-stage R&D project.
During the placement I carried out comprehensive competitor analysis, market research, and then utilised the Blue Ocean Methodology explored in the I+E module to identify the products ‘story’ and market positioning. Furthermore, I developed on my knowledge from Computing 2 with database design, account security and hashing functions. Also, I built upon my ML and Data Science knowledge from modules to create auto ML algorithms.
I implemented all features of the platform, but the contribution that was most significant was my unique search algorithm, that involved data science and SQL operations aswell as NLP for finding potential matches. This proved to be very successful and was a novel solution, making use of recent advancements in AI. Also, I developed an algorithm that was successful at merging datasets despite different formatting. Finally, I used Shapley values to drive the calculations for payment distribution, as well as designing an intuitive user interface.
