Polaris: Smarter Recruitment with AI

Design launched in 2023

B2B2C SaaS, Dashboard, Data visualization

Small to midsize companies often struggle with limited resources and tight timelines when hiring. Polaris, a Harvard Innovation Lab-incubated startup, tackles this challenge with a B2B2C SaaS platform powered by LLMs and AI to streamline recruitment and surface top talent. I lead the design of two core MVP features, driving the platform’s beta launch in October 2023 and early customer adoption in 2024.

Role

Product design intern responsible for user research, feature scoping, and iterations

Team

2 founders, 1 product manager, 2 engineers, 3 designers

Duration

16 weeks (May - August 2023)

🩵 OUTCOME HIGHLIGHT 🩵

I designed two end-to-end features, job analytic and candidate report, to help hiring teams find the top candidates faster and easier

Customize job analytics tailored to hiring needs

Quick interview setup with AI-generated keywords

Identify and learn about top candidates with insights

PROBLEM AT HAND

Hiring in tech can be a headache :(

Overwhelming volume of job applicants

Hiring teams nowadays deal with an overwhelming number of applicants, with some positions receiving over a thousand application

Easily miss out candidates if not screened thoroughly 

The current tools, like Workday, require hiring teams to manually screen each candidate, which drastically increases their workload

PROBLEM STATEMENT

How can we provide companies and recruiters with a tool to quickly screen through a large pool of applicants, surfacing top candidates they need?

RESEARCH & DISCOVER

Identify gaps in today’s hiring solutions

Existing solutions lacked advanced analysis capabilities

I did a competitor analysis on four existing products to identify gaps and opportunities.

Despite existing solutions, recruiters still struggle with two key pain points

Despite current market solutions, users still struggled inefficiency problems. I took the learnings from survey results of 52 recruiters and business owners and turned them into initial design ideas.

I deep dive to find out what recruiters need the most help with to better screen candidates

To explore these design opportunities, I led user interviews with hiring managers and recruiters to understand what do they cared the most about in resume and phone screen.

Finally - narrow down two design features for MVP

Summarizing all findings from primary and secondary research, finally, we came up with 2 design features for our product.

DECISIONS & ITERATIONS

Challenge 1. how do we help users create a tailored job analytic to save time in screening candidates? 

A customizable 3-layers keyword matching for resumes

To incorporate our learnings from user interview, we designed a 3-layer keywords description system for resume screening.

Iterate for better visibility and feature adoption

For setting up screening questions and keywords, I iterated on ways that AI can better help users to extract keywords from the interview questions they provided. I strive for intuitiveness and clarity in this design.

OUTCOME HIGHLIGHT

AI identify keywords for candidates analysis so you don't have to

Here is the final look! AI technology is used to to extract relevant keywords from interview questions, speeding up the analytic setup process and ensuring we capture the most crucial data.

Define the north star for design

Challenge 2. how could we best present the analysis result so that users can quickly find top candidates and understand them better?

Get an overview of the application pool

uSER GOAL

Need a quick snapshot of the applicant pool to decide on next steps, such as how many interviews to send or if the job description need adjustment.

kEY METRIC

Time-to-hire, Match quality

sOLUTION

A pie chart highlights overall keyword matches and average years of experience, giving a quick view of applicants quality.

See top-matching candidates first

uSER GOAL

Quickly find candidates with the most relevant experience.

KEY METRIC

Match quality

sOLUTION

The resume report ranks candidates by experience level (using 3 level requirement set up, basic, preferred, bonus) with a stacked bar chart.

Dive deeper to see best fit

uSER GOAL

Visually compare which candidates best meet preferred and bonus criteria.

KEY METRIC

Match quality

sOLUTION

A scatter plot compare candidates’ competency based on two attributes( preferred and bonus criteria) , with top-right placements indicating stronger matches.

DESIGN OUTCOME #1

Create a job analytic by going through a step-by-step keywords set up process

DESIGN oUTCOME #2

View analytic report to find the top matching candidates and see AI hiring insights

LEARNINGS

Alignment of user and business goals

Clearly identify our customers and deeply understand their needs and pain points was vital for aligning our design solutions with business goals.

Prioritization through cross-team collaboration

Navigating constraints became a key part of my design process, as it defines what we can do and what approach to take. Constraints can change throughout the project, so regular check-ins with product owners kept me agile, allowing us to adapt quickly to changes and ensure our designs stayed relevant and impactful.

See more of my work

AI-Powered essay writing tool for college applications

Design to improve user engagement in mobile E-learning