Kruthika Pinnada - 2023 Poster Contest Resources

Kruthika is a third-year student at RV College of Engineering, pursuing her studies in Computer Science and Engineering. She had joined HPCC Systems as a summer intern in 2023, where she contributed to the project - Resume Analyzer, leveraging the power of NLP++ and HPCC Systems. Kruthika's keen interest lies in working on diverse machine learning and artificial intelligence projects, reflecting her passion for cutting-edge technology and innovation.

Poster Abstract

With today's increasing digitalization, it has been observed that the vast majority of processes are becoming automated. One such occurrence that has been noted is the hiring of applicants for various positions in companies. Due to the overwhelming volume of resumes received for identical jobs and positions, recruiting has become time-consuming and challenging. Consequently, there is a growing demand for technologies that can automate this procedure so that resumes may be screened, sorted, and prioritized. Current systems rely heavily on keyword searches and thus struggle with understanding the context, resulting in misinterpretation or misrepresentation of information on the resume. Through traditional systems of analysis- if a query was to find resumes having Java as a skill, we would get results pointing to both programmers working with Java and biologists working on Java Island which leads to incorrect results. Additionally complex sentence structures, unusual formats, and industry-specific jargon can further challenge the accuracy of the analysis. The project "Resume Analyzer" leverages the power of NLP++ programming language to build a digital human reader that parses the resume text in the same we humans do. The system has made use of the “zoning” of a resume (done by a previous intern) and aims at doing an in-depth analysis of text and extracting valuable information in the way a human does. It relies on a powerful rule system as well as predefined and custom dictionaries which help isolate text of importance and also assign an attribute to certain text based on the context of use. Additionally, the parse tree is used heavily to look at the text and group it hierarchically with the view to isolate different sections of the text. After this the section-wise analysis is done and information like - Name, email, links, telephone number, dates, date ranges, bullets, prose, grades, institutions, company names, languages spoken, Programming Languages etc are extracted. All this information is put to the Knowledge Base (KB) which serves to create a comprehensive picture of the individual and also serves as a repository of information on the candidate in a hierarchical fashion. The KB is then converted into XML, which is a format very popularly used by ECL, so that a record can be created out of the KB. This record serves as a database and based on the requirement of the organization, it can be selectively queried to find out candidates having appropriate requirements. This project replicates human-like comprehension of resumes and does not use language models, but general rules of analyzing text which are almost identical to human interpretation of text. The information after being placed into an ECL Record can easily be accessed with the help of ECL Queries, thereby making the job of parsing a multitude of resumes, easier, thereby leveraging the HPCC Systems.


In this Video Recording, Kruthika provides a tour and explanation of his poster content.

Resume Analyzer

Click on the poster for a larger image.

All pages in this wiki are subject to our site usage guidelines.