VDAB, vastly improved job search user experience by applying intelligent language models. Try it on the vdab website.
At VDAB, the Flemish public employment service, the online job search is critical. With tens of thousands of jobs and thousands of job search requests every day, good search quality is key.
theMatchBox assessed their search engine and was able to significantly increase the number of relevant results by combining intelligent language models and terminology extraction based on the content of job offers.
By resolving phrasing variations (feminine/masculine/mixed, singular/plural, splitting compound words…) and fixing spelling errors in three languages (Dutch, English and French), text queries like (in Dutch) “kapster” will return the same results as “kapper”.
It was a pleasure working together with VDAB’s search & match team to integrate our technology.
Seats2Meet, improved quality of skills and knowledge entries in their users’ profiles
Seats2Meet offers co-working places all over The Netherlands and keeps profiles of all of their members, containing their skills and knowledge. theMatchBox used its language models to improve the quality of the profiles by mapping all variations of writing a skill to one single concept. Because of this, the quality of searching and matching profiles improved and it also allowed them to build an intelligent suggestion model to help users enter new skills.
FlexService, implemented a matching engine between jobs and candidates based on Elasticsearch
FlexService provides software for temp agencies, where search & match on job offers and candidates plays an important role. theMatchBox fully integrated a search & match engine based on Elasticsearch and handled the whole process, from indexing the database and integrating with the current search & match screens to training the developers and support staff.
Timing, reduced the load on the help desk by improving the knowledge base search
Timing is a large temp recruitment agency in The Netherlands. For all its offices, Timing has set up a central help desk for operational questions. In order to reduce the load on this help desk, theMatchBox improved the search on the knowledge base by combining intelligent language models and terminology extraction based on the content of their knowledge base. The search engine is built on Elasticsearch and theMatchBox handled the integration with the existing content management system.