STARTUP

LIVE

AI-ASSISTED

Reddy research and evaluation that shaped a live hiring platform

Reddy research and evaluation that shaped a live hiring platform

I audited, analyzed, and pressure-tested an early-stage recruiter product — one of my proposed features shipped in the rebrand, and my brainstorm is now live. Built with AI-assisted workflows.

Role
Product / UX Designer

Timeline
May 2026 – Present

Team
Founder + eng + design

Focus
Research, evaluation, visual design

Status
Web app live · testing next

What I owned: Heuristic evaluation, feature mapping, competitive analysis, web-app direction, cross-platform consistency audit, marketing templates, and production graphics.

01 — EVALUATE

I started by understanding the product deeply

I started by understanding the product deeply

Before proposing anything, I audited the existing app against usability heuristics and mapped its full feature set — surfacing friction points and giving the team a shared picture of what existed and where the gaps were.

Learning from the landscape

Learning from the landscape

I ran a competitive analysis across comparable platforms — comparing onboarding, filtering, and candidate management — to find the gaps Reddy could own.

02 — SHAPE & SHIP

A feature I proposed shipped in the rebrand

A feature I proposed shipped in the rebrand

The competitive analysis revealed recruiters had no structured way to line up candidates. I proposed a queue to stage and prioritize applicants without losing context. The team adopted it — and it became part of the product when Reddy was rebranded and rebuilt.

[The gap]

Gap identified →

[My queue wireframe]

Proposed direction →

SHIPPED

From brainstorm to a live interface

From brainstorm to a live interface

I led a brainstorm for the web app — mapping the recruiter’s core loop and exploring directions. A version of that thinking is now live, evolved through collaboration with the team.

[Live site]

View live →

SHIPPED

The live version of the web application

The live version of the web application

I brainstormed and came up with rough dark versions of the web application. It was tweaked in production, but the core is mine and it’s live.

03 — VALIDATE

Auditing web and mobile for consistency

Auditing web and mobile for consistency

Once the web app was built, I ran a detailed comparison against the mobile app — checking components, terminology, spacing, and interaction patterns for inconsistencies, and documenting every mismatch and its fix so the two platforms feel like one product.

[Web]

[Mobile]

Annotation space between the two artifacts for terminology, spacing, component, and interaction mismatches.

A flexible marketing template, not a one-off

A flexible marketing template, not a one-off

I built an editable marketing slide template with multiple variations, so the team can adapt messaging without redesigning from scratch — a small system rather than a single deliverable.

One template, multiple variations.

View the marketing template →

TESTING NEXT

Validating the design — testing next

Validating the design — testing next

The next step is a usability study: I’m writing the script now and will run it in the coming weeks, tracking task success, time-on-task, and errors. Findings will drive the next iteration.

Results and iterations coming soon.

AI-assisted workflow

AI-assisted workflow

Research synthesis

Clustered evaluation and competitive notes into themes faster.

Exploration & variations

Generated template and graphic directions to choose from.

Production speed

First-pass drafts I then art-directed.

I directed, validated, and owned every decision. AI accelerated the work; it didn’t make the calls.

Impact & What’s Next

Impact & What’s Next

Proposed queue feature shipped in the rebrand.

Web-app brainstorm and graphic are live.

Web/mobile consistency audit delivered.

Usability testing next — results will drive v2.

This is a boring problem to solve. That's exactly why it hasn't been solved.

This is a boring problem to solve. That's exactly why it hasn't been solved.

— Joe Chen, Reddy

Shared with permission. Some data anonymized for confidentiality.