21 Oct Aphelia – An AI based resume parser
“The potential of using AI in the recruiting process is best for large companies and could radically alter the hiring process.”
According to Glassdoor, on average, a corporate job offer attracts nearly 200 resumes, and of course, by the norm, it will be a human recruiter who has to go through all of the received 200 resumes.
Due to these high number of applicants and resume submissions to job postings, recruiters find it challenging and tedious to filter resumes. In addition, all resumes are not written in a fixed structure. For example, in some resumes, candidates present their skills in a very graphical format while others are presented in a very textual format. This increases the difficulty for the recruiter to infer and extract the right information from each resume.
Reading and understanding a CV is very much of a human task since there are thousands of different formats that cannot be easily identified by a machine. Since machines cannot read a CV like a human, the potential of automation on recruitment platforms is lower than it could be.
Many existing HRM systems do not compose a component for resume parsing extraction because getting structured data from an unstructured text is not an easy task. Most resume parsing and extraction products available in the market use rule-based or Named Entity Recognition techniques to extract and identify particular words that correspond to the attributes of the candidate. The problem with rule-based methods such as using regular expressions is that these rules aren’t flexible enough to cover all possible resume formats. In named entity recognition (NER) based resume parsers, only words can be categorized, which wouldn’t work in the case of resumes because, in resumes, different attributes are of different structures.
The complexity of solving the problem was also increased because resumes are considered sensitive personal data, so we had to resolve the bottleneck of data scarcity.
Aphelia can solve the need for extraction and structured presentation of data in existing recruitment systems so that recruiters can view each resume in one structured way instead of going through countless different structures and formats and then interpreting information from them.
Another unique feature of Aphelia is that it can get better at information extraction as it is fed more and more resumes due to the extraction algorithm’s learning capability. Also, this automated self improvisation is rarely seen in most of the similar products.
Aphelia’s market potential is not only limited to the enterprise recruitment space. It can also help other Job sites such as Indeed / Glassdoor to rank candidates based on the extracted data and give a preview of each candidate based on their resume to the recruiter.