􀄪 More projects

About

Résumé

Simplifying Mass Hiring with AI to Scale a Nonprofit's Fellowship Program

Hackathon / Case Study | Gen AI, Prototyping, User testing

􀄪 More projects

About

Résumé

Simplifying Mass Hiring with AI to Scale a Nonprofit's Fellowship Program

Hackathon / Case Study | Gen AI, Prototyping, User testing

􀄪 More projects

About

Résumé

Simplifying Mass Hiring with AI to Scale a Nonprofit's Fellowship Program

Hackathon / Case Study | Gen AI, Prototyping, User testing

Overview

Overview

Environmental Defense Fund (EDF) partnered with McKinsey Digital Hackathon to tackle a challenge: scaling up their fellowship program. In collaboration with three engineers, I designed an AI-driven recruitment solution, enabling mass hiring. Following the Hackathon, I revisited the project to explore design with GenAI, prototype polished interactions, and validate the design with users.

Environmental Defense Fund (EDF) partnered with McKinsey Digital Hackathon to tackle a challenge: scaling up their fellowship program. In collaboration with three engineers, I designed an AI-driven recruitment solution, enabling mass hiring. Following the Hackathon, I revisited the project to explore design with GenAI, prototype polished interactions, and validate the design with users.

Overview

Winner, Chosen as the best solution by EDF

Overview

Time
July 15 - 16, 2023 (24 hours),
April - May, 2024 (4 weeks)

My role
Product designer

Overview

Tool
Figma,
Lottie, Runway

Tool
Figma, Lottie, Runway

Collaborators
2 Backend engineers,
1 Frontend engineer

Overview

Problem Statement

How can EDF scale the Climate Corps fellowship program from 150 to 1000 fellows?

They have many qualified candidates but have trouble accommodating more with their existing resources.

Solution

AI-powered end-to-end recruiting platform with process inversion enabling mass hiring

Automate and streamline key hiring processes, and implement process inversion to enable mass hiring with the same resources.

Personal Session is a new TikTok feature that provide purpose and end point to passive users with niche themed viewing sessions. The sessions end with engaging activities to help users to lea smoothly.

Design Solution

01

Simplified Mass Screening & Matching with AI

EDF currently spends 3 weeks with 13 recruiters just screening candidates. Recruiters can effortlessly screen 1000+ applicants with AI scores and then match qualified candidates to organizations. Yet, they can still review and grade profiles manually, but AI insights speed things up.

02

Streamlined Specific Low-Volume Candidates Tasks with Conversational AI

Following the initial high-volume screening, recruiters can use conversational AI to handle low-volume tasks. They can easily type and ask specific requests, such as "find candidates with 2+ years of data science experience for the Circular Economy Fellow role".

03

Seamless Communication with Collaboration Bar

During interviews, recruiters highlighted the challenges of communicating and collaborating with contract recruiters. The collaboration bar addresses this by allowing users to easily view team members' tasks and their progress.

Initial Process

Hackathon (24 hours)

Hackathon (24 hours)

Hackathon

Within 24 hours, participants were asked to present a solution for McKinsey's partnering nonprofits, and my team was matched with EDF.

Within 24 hours, participants were asked to present a solution for McKinsey's partnering nonprofits, and my team was matched with EDF.

-

Problem Statement

How can EDF scale the Climate Corps fellowship program from 150 to 1000 fellows?

What is Climate Corps?

Environmental Defense Fund is a nonprofit addressing climate change.

Climate Corps is a summer fellowship program that matches grad students with companies to address sustainability challenges.

Identifying
Problem space

Identifying Problem space

Our team concluded that the best way to scale the fellowship program is to accept more applicants. We identified two problems that prevented EDF from achieving this.

Via interviewing stakeholders and reviewing the provided document.

Our team concluded that the best way to scale the fellowship program is to accept more applicants. We identified two problems that prevented EDF from achieving this.

Via interviewing stakeholders and reviewing the provided document.

-

Solution

We prosed a solution that addressed both problems and improved EDF's hiring efficiency by 89%. It was chosen as the best proposal by the nonprofit.

We prosed a solution that addressed both problems and improved EDF's hiring efficiency by 89%. It was chosen as the best proposal by the nonprofit.

Solution Proposal

AI-powered end-to-end recruiting platform with process inversion enabling mass hiring

*Assumptions: resume screening (25mins), interview (1hr), matching (45mins)

-

Refinement Process

Revisit (4 weeks)

Revisit (4 weeks)

Revisit

I revisited the project personally to learn more about designing AI interfaces, prototyping polished interactions, and validating assumptions through user testing.

I revisited the project personally to learn more about designing AI interfaces, prototyping polished interactions, and validating assumptions through user testing.

01
Research

For the first step, I conducted research to understand AI usage and user needs in the recruitment context. With these new insights, I set three goals for the redesign.

Method: Landscape research, interview w/ an industry professional, revisit hackathon note

For the first step, I conducted research to understand AI usage and user needs in the recruitment context. With these new insights, I set three goals for the redesign.

Method: Landscape research, interview w/ an industry professional, revisit hackathon note

-

01

Enhance the human touch in the AI screening/matching process.

02

Leverage conversational AI for low-volume candidates processing.

03

Introduce a collaborative feature to encourage seamless communication with fellow recruiters.

02
Wireframe

Then, I rapidly iterated on wireframes to ideate new features and build out enhanced flows and interactions.

Explorations and iterations include:
Landing screen, AI-generated loading screens, dynamic AI input headers, AI highlights on candidate profiles, and more.

Then, I rapidly iterated on wireframes to ideate new features and build out enhanced flows and interactions.

Explorations and iterations include:
Landing screen, AI-generated loading screens, dynamic AI input headers, AI highlights on candidate profiles, and more.

03
Visual Design

To add a human touch to the AI experience, I explored various ways to introduce clear visual language.

To add a human touch to the AI experience, I explored various ways to introduce clear visual language.

AI features
visual Assets

AI features visual Assets

I used a spark logo, designated color, and a gradient to clearly indicate AI features throughout the platform.

illustrations

To further help users understand each hiring process, I AI-generated illustrations, tested variations, and incorporated them into the design.

04
User Testing

04
User Testing & Refinement

I conducted a usability test with three individuals to test the clarity of each flow and the users' understanding and engagement with the AI features.

I conducted a usability test with three individuals to test the clarity of each flow and the users' understanding and engagement with the AI features.

Plan

I planned clear objectives to achieve actionable insights for improvement

-

Findings and Iterations

I refined the designs to address the recurring user feedback.

05
Next Step & Reflection

01

Onboarding with a thorough feature walkthrough guide

I learned commonly used AI interfaces and copies can be interpreted differently in the recruitment context. More robust guidance, such as a feature walkthrough or information tooltips, could help users navigate the platform more easily.

Onboarding with a thorough feature walkthrough guide

I learned commonly used AI interfaces and copies can be interpreted differently in the recruitment context. More robust guidance, such as a feature walkthrough or information tooltips, could help users navigate the platform more easily.

02

Design Humane AI experience

One of the biggest challenges of this project is bringing the backend solution into the user-facing experience. This allowed me to explore and learn about improving AI-human interaction through various tactics such as lagging loading/generating times to deliver informative messages,

Design Humane AI experience

One of the biggest challenges of this project is bringing the backend solution into the user-facing experience. This allowed me to explore and learn about improving AI-human interaction through various tactics such as lagging loading/generating times to deliver informative messages,

Design system for a social giving platform 􀄫

Design system 􀄫

ⓒ 2024 Junhyung Cho

Email

Linkedin