Challenge
Recruiters spend a lot of time every day reviewing resumes to find potential candidates, contacting them, inviting them for interviews, and forwarding resumes to the hiring department. These daily tasks involve many repetitive and time-consuming activities.
Outcome
Solution
Daily Recommendation Algorithm:
Enhancing recommendation accuracy by prioritizing the selection preferences of Recruiters.
In the candidate selecting process, recruiters hold various requirements on different aspects. In other words, the matching between candidates and occupations is decided by different weight. Therefore, we designed the three hierarchy filtering and sorting mechanism to select matching candidates based on the keywords extract from the job description.
STAGE 1: involves accurately screening out candidates who are not suitable in terms of location and position.
STAGE 2: candidates are filtered flexibly based on their movements as well as the preferences of recruiters.
STAGE 3: focuses on the matching rate, which allows the mechanism to be more flexible. Candidates are scored based on several relevant factors, such as work experience, previous industry, and salary expectations. For example, a new grad from Harvard may get 1 point in work experience but 5 points in education.
After the algorithm's launch, the click rates for the resume recommendation module were boosted by 22%.
Intelligent Calling System
Built quicker, timely communication channel between recruiters and candidates
Due to the fact that the resumes in our recruitment website are not uploaded by candidates themselves, we need to consider the issue of candidate privacy protection.
After finding that SMS inquiries about job status were less efficient, we designed an intelligent phone system to connect recruiters with candidates.
After the launch, the response rate increased by 43%, and the operation of recruiters' invitation to jobseekers increased by about 3 times.
Reflection
Other projects
Ceribell: Medical Notification System Research & Design
I contributed to designing the medical notification system to improve communication in healthcare institutions, resulting in a user satisfaction score of 8.6 in the pilot version.
ClipEase: Smart Clipper User Research
I assisted a Chinese AI-driven haircutting solution provider in exploring the U.S. market through in-depth interviews, sidewalk surveys, and quantitative analysis.
WriteWise: Elevating English Writing Concept Design
I created a website concept design utilizing a large language model to assist non-native speakers in improving their writing skills.