Jingyi Zhang


> Index
> Work
> Project

> About

I design AI-powered product with creativity and systematic thinking.

B2B/B2C SaaS for creative & production workflow.

Currently @ReelShort.




© 2026
       
         © 2026

  002  
ReelShort Casting System — Casting at 100M+ View Scale



Reelshort Casting System

1
111



Redesigned and launched an internal casting platform for the leading global short- drama company, featuring a talent database and project management dashboard to support faster production decisions across high-volume TV productions.

ROLE

PRODUCT DESIGNER
BRAND DESIGNER
SCOPE

Product  Design
User  Experience Research
Design Engineer


Confidentiality note: This is an internal ReelShort tool. All screens, names, photos and figures in this case study are fictional demo data or have been anonymized. No real talent, agency, contract or production information is shown.  

2026-10-01-10.05.26



 Overview  

ReelShort produces a high volume of short-form dramas every year, which means casting teams constantly need to find, evaluate, and coordinate with a large pool of actors.

The existing Casting Hub helped centralize this process, but finding the right talent still required navigating dense profiles, incomplete information, multiple platforms, and repetitive communication.

I redesigned the platform around one goal: Make casting decisions faster with less manual work.

2026-09-29-9.28.33
Reelshort App on IOS received more than 500K ratings

 
2026-09-29-4.45.54
Reelshort Produce more than 1000+ short drama series every year




  The Problem  

Imagine you are a producer/casting director casting a new short drama.You need to search for actors across IMDb, Instagram, agencies, and internal records.  Then you compare photos, check availability and location, organize contact information, reach out, and track responses.

A lot of time is spent simply collecting information before a casting decision can even happen.The opportunity was to turn the Casting Hub from a talent database into a more useful decision-making tool.


Casting_Chaos




  User Research  

Before redesigning the platform, I wanted to understand what casting professionals actually pay attention to. I did research with five Producers and Casting Directors working across short-form entertainment and fashion about how they find talent, compare candidates, and make final decisions.

I spoke with five Producers and Casting Directors working across short-form entertainment and fashion. Four needs came up repeatedly: Visual Fit, Complete Information, Chemistry, Agency Network. These insights became the foundation of the redesign.

Casting_Needs_HD




  Key Design Decisions  

All screens in this section were rebuilt with fictional demo data for this portfolio. Actor names, photos, measurements, agency names and contact details are invented or generic placeholders. Real talent data is confidential and is never displayed here.

  01 · Make Talent Card More Visual  

Multiple photos can be viewed directly from the card, while the information hierarchy prioritizes the details needed for quick decisions. Casting teams can screen talent from the grid and open fewer profiles along the way.

card111_HD
0930_HD


Talent-Card--Problems_
Problems
The old card showed only one photo, so users kept opening profiles to see other angles. Its tags were cluttered, and empty values like "AI Unknown" pushed the useful details out of sight.


Talent-Card--Case-Study_
What changed
  1. Photo carousel. Users flip through every photo right on the card with the ← → arrows.
  2. Simplified tags. Only filled values are shown, with short, colour-coded labels like "Int 4".
  3. Profile completeness. A colour-coded score next to the name shows which profiles still need data.

   

  

 02 · Surface Information Completeness   

I identified 12 key data points commonly needed during casting and introduced a Profile Completeness indicator. Casting teams can quickly see which profiles are ready for consideration and which are missing critical information.

Profile-Completeness
Profile Completeness Setting System


Frame-1
Profile Completeness UI Component



  03 · Rebuild Agency Library Page  

I redesigned the Agency Library to connect agencies directly with the actors they represent.  Casting teams can move between an agency, its contacts, and its talent pool without treating them as separate databases.


Agency-Table--Problems
Problems
  1. Every cell was an input. Long names and emails got cut off, and rows were hard to read.
  2. Loud, unclear status. Green emoji checkmarks filled every column, and "—" could mean no or unknown.
  3. Missing data blended in. Empty fields looked the same as filled ones.
  4. Delete on every row. One click could remove an agency record.


Agency-Directory--Redesign-Demo-data
What changed
  1. Add Agency. One clear button adds agencies, and each new record needs at least one public profile.
  2. Richer agency info. Each row shows office locations and Website, Instagram and LinkedIn links.
  3. Cleaner status. Soft ✓ pills mark what's done, and grey "—" pills mark what isn't.
  4. Actors with faces. Linked actors show as a face stack with a count that opens the roster in one click.



  04 · Change the Filter’s Information Architecture  

The original filter offered comprehensive options, but the structure was cluttered and single-select interactions made complex searches inefficient.
I reorganized the hierarchy, merged overlapping criteria, removed unnecessary options, and introduced multi-select filtering.A cleaner, more flexible filtering experience helps casting teams narrow down talent faster.


Library_Filters_Problems-1


Library_Filters_WhatChanged_1-1_
What changed (overview)
  • ① Simpler structure (bracket on the left): Six rows became three: library tabs, one search bar with sort, and seven grouped filters.
  • ② Multi-select (dashed box around the filter buttons): Each filter group opens a panel with checkboxes.
  • ③ Logical order (labels under the buttons): Who → Look → Where → Deal → Consent: who the person is, how they look, where they're based, how we work with them, and what they've agreed to.


Library_Filters_WhatChanged_2-1
Inside the filters (panel details)
  • Gender: Pick any mix of genders. Minors is now its own Exclude / Include / Only switch, because it's an age, not a gender (② ③).
  • Age: Age ranges calculated from date of birth, and you can tick more than one (②).
  • Appearance: Height, hair and eye colour are in one panel instead of three separate dropdowns (①).
  • Consent: AI friendly and minimum intimacy level are grouped together (①).



  05 · VFX Information UI  System’s  Update  

The VFX grading standard is the reference the team uses to grade every project's visual effects, from D (simple shot) to S+ (top-tier difficulty).  
I turned the old full-width table into a collapsible card with colour-coded grade chips, a lighter table and a short pricing footnote. The project list now sits near the top of the VFX page, and the full standards are still one click away.


2026-09-30-8.36.15


VFX_Grading_Redesign-3
What Change:
  1. Collapsible reference: the standards now sit behind a single summary card ("D → S+ · …"), and the full table only opens when you click View standards.
  2. Grade chips: a row of chips lets you scan D → S+ at a glance, with colour tiers for the higher grades. The A grades get a dark green chip and the S grades an amber one.
  3. Cleaner table: heavy cell borders are replaced by row dividers only, with small uppercase headers and more spacing.
  4. Pricing as a footnote: the batch discount text moved from above the table to a short note below it, so the grades stay the focus.



 Outcome  

Casting looks like picking faces, but most of it is keeping track of information. One role can mean hundreds of actors, each with photos, measurements, an agency, consent limits and a schedule. Before the ReelShort Casting Hub, all of that lived in spreadsheets, inboxes and people's memory. I designed the hub to bring it into one place.
The hub now holds 600+ actors, 130+ agencies and 60+ projects, and it has become part of a larger internal system. It didn't replace casting judgment. It put everything the team needs in front of them.
Figures are rounded for confidentiality.

Outcome_Illustrations


Outcome_Icons