HR & Job Search
Vector search engine can be used to match candidates and jobs even if there are no matching keywords or explicit skill descriptions. For example, it can automatically map ‘frontend engineer’ to ‘web developer’, no need for any predefined categorization. Neural job matching is used at MoBerries for automatic job recommendations.
There are multiple ways to discover things, text search is not the only one. In the case of food, people rely more on appearance than description and ingredients. So why not let people choose their next lunch by its appearance, even if they don’t know the name of the dish? We made a demo to showcase this approach.
Increase your online basket size and revenue with the AI-powered search. No need in manually assembled synonym lists, neural networks get the context better. With neural approach the search results could be not only precise, but also personalized. And Qdrant will be the backbone of this search.
Read more about Deep Learning-based Product Recommendations in the paper by The Home Depot.
User interests cannot be described with rules, and that’s where neural networks come in. Qdrant vector database will allow sufficient flexibility in neural network recommendations so that each user sees only the relevant ad. Advanced filtering mechanisms, such as geo-location, do not compromise on speed and accuracy, which is especially important for online advertising.
Not only totalitarian states use facial recognition. With this technology, you can also improve the user experience and simplify authentication. Make it possible to pay without a credit card and buy in the store without cashiers. And the scalable face recognition technology is based on vector search, which is what Qdrant provides. Some of the many articles on the topic of Face Recognition and Speaker Recognition.
Customer Support and Sales Optimization
Current advances in NLP can reduce the retinue work of customer service by up to 80 percent. No more answering the same questions over and over again. A chatbot will do that, and people can focus on complex problems. But not only automated answering, it is also possible to control the quality of the department and automatically identify flaws in conversations. Read more about the “Sentence Embeddings for Customer Support” case study.
Empower shoppers to find the items they want by uploading any image or browsing through a gallery instead of searching with keywords. A visual similarity search helps solve this problem. And with the advanced filters that Qdrant provides, you can be sure to have the right size in stock for the jacket the user finds.
Fraud detection is like recommendations in reverse. One way to solve the problem is to look for similar cheating behaviors. But often this is not enough and manual rules come into play. Qdrant vector database allows you to combine both approaches because it provides a way to filter the result using arbitrary conditions. And all this can happen in the time till the client takes his hand off the terminal. Here is some related research paper.
Law Case Search
The wording of court decisions can be difficult not only for ordinary people, but sometimes for the lawyers themselves. It is rare to find words that exactly match a similar precedent. That’s where AI, which has seen hundreds of thousands of court decisions and can compare them using vector similarity search engine, can help. Here is some related research.
Media and Games
Personalized recommendations for music, movies, games, and other entertainment content are also some sort of search. Except the query in it is not a text string, but user preferences and past experience. And with Qdrant, user preference vectors can be updated in real-time, no need to deploy a MapReduce cluster. Read more about “Metric Learning Recommendation System”
The growing volume of data and the increasing interest in the topic of health care is creating products to help doctors with diagnostics. One such product might be a search for similar cases in an ever-expanding database of patient histories. Search not only by symptom description, but also by data from, for example, MRI machines. Vector Search is applied even here.