In this post, I’ll share some details on how we built a small application that matches resumes (CVs) with active job listings in Denmark. We use this as an internal tool, and we have recently soft-launched a customer facing product with the same functionality at its core.

Recently, I gave a talk at Solita Summit 2023. The topic for the summit was Generative AI and the other speakers were Google CTO Ted Schönbeck, and Kela CIO Janne Pulkkinen. All talks revolved around adoption of Generative AI and going beyond the hype and focusing on actual use cases and solving business problems.

Organizations are moving data workloads to the cloud, in part because of the enormous demand for compute by larger machine learning models, - and even larger data sets. For us, it was the sum of both minor (and some critical) issues with our legacy data platform, that prompted us to finally make the move and migrate everything to our new cloud data platform.

Let’s say all you know about your customers is their purchase history. With this data you want to generate individual recommendations representing cross-selling opportunities. How? Association Rule Learning.

So, you’ve decided you want to build a recommendation system where the item you want to recommend is a text-based document of some sort. This could be an article, a peer-reviewed journal entry, a series of blog posts, or perhaps an event-description or similar.