Foreword
The services of PRAXI Human Resources are aimed at identifying, on the one hand, managerial figures with top organizational positions and responsibilities, and, on the other hand, extremely specific professionalism regardless of classification. For these types of searches, the company has always used SelPer, the tailor-made management software designed and implemented by PRAXI Digital Solutions.
With continuous development supported by Scrum methodology, SelPer enables it to cover the evolving needs of recruitment process management in all its steps, from announcements to applications, profile identification and interview management.
One of the implementation requests that emerged in 2020 was in response to the need to search for keywords within candidates’ resumes in a “google-like” manner, so as to obtain a shortlist of profiles that would best cross-reference the characteristics of specific open positions.
Project
SelPer, which already allowed rich and refined searching by structured fields in 2020, has thus been enriched with a new feature of free full-text searching within hundreds of thousands of stored and classified curricula and related documents, thanks to the adoption of ElasticSearch technology, an “open source enterprise search engine” developed in Java. ElasticSearch is categorized as a NoSQL database, in that storage is not structured (with classic text/data/number fields) but dynamic, and SQL cannot be used for queries.
The main objective of this kind of search engine is to ensure optimal performance in full-text searches by indexing, i.e., by transferring the original data into a specially designed repository for the engine’s use and analyzing its contents: during a search, instead of examining all the documents, the index is checked, with a relevant time reduction. Indexing of documents takes place almost in real time and, therefore, the data contained in SelPer’s main database and the data indexed in ElasticSearch can be considered identical and aligned (a resume just entered by a candidate immediately falls under the investigations carried out by PRAXI HR consultants).
In addition to allowing full-text searches within the candidate’s professional characteristics and geo-location from the CV, ElasticSearch technology is also capable of specifying the maximum geographic distance between the candidate’s location and what will be the job site, so as to highlight a cluster of profiles within a certain mile radius.
Results
The benefit for PRAXI Human Resources consultants who use SelPer management on a daily basis concerns both the quality of the profiles that emerge from the new type of search and the speed with which the results are obtained.
In addition to indications of resumes that respond to freely entered queries, the output also includes a ranking of profiles, associating a higher score with those that best meet certain search parameters. For example, if a curriculum includes a certain lemma that is considered core, it will float in top positions.
At a later stage of release, the ability to find CVs in which two words appeared with a maximum distance between them was also added. Searching, for example, for the two words “GDO marketing,” the new feature allowed for a different score depending on whether the two words searched were far apart (“marketing ma-nager with specialization in the digital world and some experience in the GDO sector”) or close to each other (“GDO marketing with 10 years of managerial experience”).
Future
In the aftermath of the implementation of the ElasticSearch engine in the PRAXI Human Resources area, with the same technological base, a new search functionality was developed for the PROPTECH management system in the PRAXI Valuations & Advisory area, where searches consist of probing thousands of documents such as appraisals, estimates, visas and notarial deeds in order to find documents related to real estate consistent with specific search criteria.
Such software, based on ElasticSearch technology, can similarly be developed for all areas where searching within a document repository with heterogeneous origins is part of the work process.
