Roman Grebennikov@shutty

Recommendations cold-start problem is not actually a problem, if you leverage content and item metadata to build your recommendations. To showcase this idea we build a movie recommender, so you can visually see the difference between collaborative-filtering and content recommendations. We made two PRs to an existing open-source project Metarank: * support semantic recommendations with cohere-ai and sentence-transformers embeddings * use qdrant as a vector search engine to quickly perform vector similarity search With these two PRs merged building such a recommender is just a matter of a few lines of YAML code. But the semantic-similarity approach is not only about movies, but can be applied more generically in traditional places like e-commerce. For example, in fashion with high inventory churn, being able to recommend something for new clothes having zero feedback is really valuable.

🗓️ Take part in this 7-day virtual hackathon from March 10 to March 17! 💻 Create AI applications utilizing Cohere's LLM-powered Multilingual Text Understanding model and Qdrant's vector search engine. ✔️ Are you new to AI or an experienced data scientist? Designer, or business developer? Regardless of your experience and background, we welcome you and value your domain expertise. 🐱‍💻 Join us for free and let's get started!