Polaris Jobs
UX DESIGN
WEB DESIGN
RESEARCH
CONCEPT CREATION
WIREFRAMING
VISUAL SYSTEM
EXPERIENCE DESIGN
PROTOTYPING
Client: PolarisJobs
Role: UX Designer-Research, Interaction Design, Visual Design, Prototyping
Team: 5 UX Designers + 1UX Researcher
Type: Desktop Web
Tool: Figma
Timeline: June 2023 - August 2023
PROBLEM
The cessation of Amazon's AI hiring tool in 2018 due to its bias against women highlighted a critical issue: AI systems can unintentionally perpetuate gender disparities in the workplace. As a female designer, this revelation resonated with me, underscoring the importance of addressing AI bias.
INSIGHT
This event ignited my interest in the evolving perceptions of AI's potential and biases, especially in the context of hiring. I wanted to delve deeper into how AI can be harnessed to promote fairness rather than exacerbate existing inequalities.
SOLUTION
Motivated by my passion for design and commitment to fairness, I joined Polaris Jobs. Our mission is to develop a proprietary AI-driven hiring tool that matches and assesses candidates based on their preferences and core competencies. By focusing on reducing biases, Polaris Jobs aims to create a more inclusive and equitable hiring process.
Challenge—
A 2017 survey reveals that 74% of organizations acknowledged the regrettable practice of hiring individuals who ultimately proved to be mismatched for their respective roles.
How do we optimize AI for a better recruiting experience?
Research
01.Competitor Analysis
In the initial phase of the engagement, my team and I spent time studying hiring platforms in competing spaces and what users are saying about the platforms. We looked at direct competitors and also did literature review.
02. Interview Outcomes
We conducted eight interviews with SME recruiters and CEO to obtain in-depth qualitative responses.
03. User Persona
I fabricated a user persona for our target users. George is a 35-year-old well-experienced HR manager in a medium tech company. Staying efficient and high quality are always his priorities in the work.
04. Key Insights
We concluded three key insights where users face frustrations and time-consumptions in hiring.
Minimizing AI biases becomes one of the top priorities.
Problem Statement——
How might we design an AI recruiting platform that
writes precise job descriptions, effectively shortlists candidates, and minimizes AI bias to save users' time and costs?
User flow
Solution
01 Login/Onboarding
Log in and create a unique company profile.
02 Job Posting
AI-recommended precise job descriptions.
03 Shortlisting Candidates
Matched talents will come to you.
04 Scheduling Interview
Automated recommend the best time to meet.
Address AI Bias —
Dive deeper into the iteration, we found,
general distrust of AI stems from high expectations.
Things become unacceptable when AI makes wrong...
01 Provide Explanation
Provide clear and proper explanations and understandings of how AI calculates scores.
02 Collect Feedback
Collect direct feedback to refine algorithms for future recommendations.
03 Pre-Screening Users
Set pre-screening questions provideding more data to analysize.
Design system
Design system
01. Noteworthy highlights
Overall, it was a fantastic experience getting to work with new teammates on this project. It was interesting to reflect on the vast amount of variables and design considerations that need to be made even in common user experiences like filling out a form.
02. What I could've done better