back to works
Talentora AI recruiting platform on desktop and mobile
Role
Co-Founder & Head of Design
Timeline
Sept 2024 – Sept 2025
Platform
Web Application
Status
Deactivated
Team
Ben Gardiner (CEO), Vincent Li (Co-founder), Lucas Yoon (Co-Founder), Seung Min-Cho (Engineer), Mehmet Battal (Engineer), Heather Davies (Designer)

Overview

Designing an AI recruiting platform to streamline early-stage candidate screening

Talentora was an AI-powered recruiting platform designed to help hiring teams evaluate candidates before traditional recruiter interviews.

Talentora platform overview across recruiter and candidate workflows

My Role

I co-founded the product and led design across the candidate and recruiter experiences. Talentora was comprised of 5 engineers and 2 UX designers when the team first formed. We worked cross-functionally every week with the help of mentors helping us take the platform from concept to a working product.

The Challenge

Recruiters spend significant time screening candidates before they ever reach an interview

Early-stage recruiting often requires teams to review large applicant pools before deciding who deserves a deeper conversation. For hiring teams, that creates both a time and resource cost while still providing limited information beyond a resume.

Talentora explored how conversational AI could help recruiters gather richer candidate information earlier in the hiring process without replacing recruiter decision-making.

The design challenge

How might we help recruiters screen candidates more efficiently with AI while still giving them enough context and control to make the final decision?

Early-stage recruiting challenge: cost vs. reliability quadrant comparing hiring tools

The Solution

We designed an AI-assisted screening experience for both recruiters and candidates

Talentora used conversational AI interview agents to conduct initial candidate assessments and organize the resulting information for recruiters. The experience connected three parts of the hiring workflow:

  1. Recruiters configure the interview
  2. Candidates complete an AI interview
  3. Recruiters review detailed candidate insights

This gave us two very different user experiences to design: a candidate-facing interview environment and a recruiter-facing system for configuring and reviewing those interviews.

Candidate experience

Talentora account invitation email for candidates joining the platformAI candidate assessment interview with video, transcript, and session controls

Recruiter experience

Talentora recruiter login and two-factor authentication screenRecruiter applicant profile with assessment scores and schedule AI interview action

01 / Recruiter workflow

Making AI interviews configurable for different hiring needs

Recruiters needed control over what candidates were being evaluated on rather than relying on a completely automated interview process.

We designed an interview setup experience where hiring teams could configure assessments around the position and determine what the AI interviewer should evaluate.

This established the recruiter as the person defining the hiring criteria while AI supported the screening process.

02 / Candidate experience

Making an AI interview feel structured without making it feel robotic

For candidates, the experience needed to communicate clearly what was happening before they entered an unfamiliar AI-led interview.

I designed the candidate journey from onboarding and interview preparation through the conversational interview itself.

The experience focused on giving candidates clear expectations while keeping the interaction simple enough that they could concentrate on their responses rather than the interface.

03 / Candidate insights

Turning AI analysis into information recruiters could actually review

One of the largest design challenges was deciding how to present AI-generated information after an interview.

Rather than treating the AI output as a final hiring decision, we designed the recruiter dashboard around reviewing candidate performance, responses, and supporting interview insights.

The goal was to make AI analysis useful as decision support while preserving recruiter judgment.

Design Process

Building the product alongside an engineering team

Talentora evolved over several months through continuous collaboration between design, engineering, and business.

Because we were building an AI product rather than a static prototype, design decisions also had to account for technical constraints, changing model capabilities, and how generated information would appear in the interface.

I designed much of the recruiter dashboard, candidate assessment and reporting experience, interview configuration, candidate onboarding, practice experience, and AI interview flow.

Branding

Talentora branding: logo, color palette, and typography exploration

Lo fi

Low-fidelity Talentora screen 1
Low-fidelity Talentora screen 2
Low-fidelity Talentora screen 3
Low-fidelity Talentora screen 4
Low-fidelity Talentora screen 5
Low-fidelity Talentora screen 6

Mid fi

Mid-fidelity Talentora screen 1
Mid-fidelity Talentora screen 2
Mid-fidelity Talentora screen 3
Mid-fidelity Talentora screen 4
Mid-fidelity Talentora screen 5
Mid-fidelity Talentora screen 6
Mid-fidelity Talentora screen 7
Mid-fidelity Talentora screen 8
Mid-fidelity Talentora screen 9
Mid-fidelity Talentora screen 10

High fi

High-fidelity Talentora screen 1
High-fidelity Talentora screen 2
High-fidelity Talentora screen 3
High-fidelity Talentora screen 4
High-fidelity Talentora screen 5
High-fidelity Talentora screen 6
High-fidelity Talentora screen 7
High-fidelity Talentora screen 8
High-fidelity Talentora screen 9
High-fidelity Talentora screen 10
High-fidelity Talentora screen 11
High-fidelity Talentora screen 12
High-fidelity Talentora screen 13
High-fidelity Talentora screen 14
High-fidelity Talentora screen 15
High-fidelity Talentora screen 16
High-fidelity Talentora screen 17
High-fidelity Talentora screen 18
High-fidelity Talentora screen 19
High-fidelity Talentora screen 20
High-fidelity Talentora screen 21
High-fidelity Talentora screen 22
High-fidelity Talentora screen 23
High-fidelity Talentora screen 24
High-fidelity Talentora screen 25
High-fidelity Talentora screen 26
High-fidelity Talentora screen 27
High-fidelity Talentora screen 28

Outcome

Taking Talentora from an idea to a working AI recruiting product

Over the course of the project, our team developed Talentora from an early concept into a functioning AI recruiting platform.

The project

  • Raised $1K in funding
  • Joined Innovate@BU's Innovation Pathway
  • Won the Audience Choice Award at BU Spark! Demo Day
  • Produced a working platform spanning recruiter and candidate workflows
Talentora team holding Audience Choice Award certificates at BU Spark! Demo Day

The experience became one of my earliest opportunities to design an AI-native product with engineers and think beyond individual screens toward the larger product and business system.

AI Ethics & Human Oversight

Deciding where automation should stop

As Talentora developed, our team began questioning how much influence AI should have over a hiring decision.

Early versions of the product placed significant emphasis on AI-generated candidate scores across areas such as technical performance, sentiment, and overall interview performance. While these scores made candidates easier to compare, they also raised concerns about giving an automated system too much authority in a process that can directly affect someone's employment opportunities.

Through discussions with our team and HR mentor, we decided that AI should support the recruiter's decision rather than make the decision itself.

We therefore shifted the recruiter experience toward providing more supporting evidence, including interview summaries, individual responses, transcripts, category-level analysis, and access to the original interview. Recruiters could use the AI analysis to identify where to look while still reviewing the underlying information themselves.

We also discussed unresolved concerns around sentiment, facial-expression, and speech analysis. These signals could potentially disadvantage candidates whose communication patterns differ because of disability or other individual differences. We did not fully resolve these questions before the project ended, but they changed how I thought about designing AI products in high-impact contexts.