About the Project
This project was inspired by our own experience as students searching for internships and graduate roles across multiple job platforms. Job information is often scattered, inconsistent, and difficult to compare with one’s actual skills. We wanted to build a system that not only aggregates job listings, but also helps students understand how well a job matches their technical background.
Through this project, we learned how to clean and standardize heterogeneous real-world datasets, especially when dealing with missing fields, inconsistent salary formats, and unstructured text. We also gained experience in skill extraction, resume-to-job matching, and building an end-to-end interactive web prototype that integrates data processing with user-facing features.
The project was built by first preprocessing six job datasets into a unified JSON format, followed by constructing skill and company vocabularies to support accurate filtering and matching. We then implemented a web-based interface that allows users to upload resumes, browse and filter jobs, compare job requirements with their own skills, and track application deadlines.
One major challenge was handling noisy and inconsistent data, particularly salary information and company naming variations. Designing meaningful skill matching without relying on external APIs was also challenging. Addressing these issues required careful rule design, iterative testing, and balancing accuracy with usability in a working prototype.
Built With
- api
- javascript
- jupyter
- python
- vscode
Log in or sign up for Devpost to join the conversation.