Microsoft's Vector Search engine uses deep learning models to represent data as vectors. The Vector Search website is an interface that demonstrates the power of these models. Users can upload an image, or choose one from a pre-generated grid of images and send it to the Vector Search API. The API will respond with images that fulfil the 'closest neighbour' criteria of the image submitted.
For this project, I coded the responsive layout and UI of the VectorSearch website. This includes the markup and styling, the routing and the interactive elements (tabs, buttons and scrolling). The responsive layout was particularly challenging as it had to be fully responsive and accommodate images generated by the AI which came in different dimensions.