CV

Senior Product Designer · B2B SaaS · Design system · Internal tools · IPTV/OTT

A design system that cleaned up 10 years of legacy

DREAMPlatform is a corporate digital TV platform (IPTV/OTT) from GS Group: operators use it to manage content, subscriptions, advertising and device updates for millions of subscribers. Its clients include Tricolor, the largest digital TV operator in Russia. I joined as the first designer on the platform and built a design system from scratch for six services.

6
platform services
40+
components with documentation
3
development teams
1 year
from start to one unified system
The DREAMPlatform library component grid: Button, Combobox, Status, Toast, Filter, Tag, Search, Calendar

Problem

10 years without a designer: what it looked like

The platform consists of six services, each with its own admin panel for operators.

MDS — metadata and the content showcase

DRM — access and subscription management

UMS — device updates

CS — entry point and request balancing

DREASYS — the advertising system

App Boutique — the app store

For about 10 years, these panels were built by different teams on Bootstrap, with no designer and no UX research. Features piled up but were never reviewed: operators worked in interfaces where some elements were unclear or unnecessary to them. In MDS forms, they kept skipping fields — they did not understand what they were for and were afraid to enter wrong data.

GS Group designers were busy with other products — for all these years, DREAMPlatform lived on development alone.

Goal. Bring six services to one interface and make support simpler — that is how it was set. I defined it more broadly: along with unification, clean out the dead functionality, set a process so that only what is needed gets into the system, and build components so that three development teams would use them, not bypass them.

My contribution — three decisions out of many made over a year of work

Decision 1. Legacy

A design system as a legacy cleanup, not cosmetics

Context. I went to the Tricolor office several times, sat next to operators and watched how they work. And I saw that the problem was deeper: it was not about how the buttons look, but about the fact that a large part of the functionality was dead.

Decision. There were two paths: limit the work to a visual update, or clean up the legacy at the same time. I proposed the second — the team supported it. Together with analysts, I defined what is really used: only what is needed went into the design system, and the rest was removed. We removed dead fields from forms. Some buttons had been disabled by development itself “just in case” — we found out it was legacy that had never been used, and removed them. Forms became shorter, and operators stopped stumbling.

ResultThe design system became a product filter. Not a visual update, but a rethink of what should be in the product.

A conscious trade-off. A full redesign of the UX flows of 6 services was unrealistic in a year with a team of 2 designers. We limited the work to a legacy cleanup and unification, on purpose, without redesigning scenarios.

Before
Before: MDS Account Manager, editing a user — a full form with extra fields
After
After: MDS Account Manager, editing a user — some fields removed
Before
Before: MDS Account Manager, the user list
After
After: MDS Account Manager, the user list in the unified system

Decision 2. Process

The design system changed the development culture

Context. Each team lived on its own: an operator asks — development builds. That is how chaos piled up.

Decision. I set a new process as a standard: any request for a new feature goes through research and approval. I defined the criteria for what goes into the system and what does not. Decisions were made together with the second designer and development, but I led the criteria for the system’s contents and the product logic. Every component went through approval before release.

ResultAdding something “because an operator asked” no longer works. The system became a coordination point between 3 teams that had not worked together before.

Process diagram: an operator request goes through research, analytics, the designer’s decision, a demo for team leads, and the rollout cycle — design, layout, review, test, release

Decision 3. Architecture

A conscious simplification for the sake of adoption

Context. The development teams had different stacks and were used to solving UI on their own. If components are complex or inflexible, teams will bypass the system and keep doing things their own way.

Decision. I made components easy to assemble and reuse: minimal nesting, universal. One system of colors, typography and spacing for all services. 40+ components with documentation in Confluence — for each one, I described where and how it is used. I applied the system as a pilot on one service: we released it, I watched operators work with the updated service, the teams accepted it — and only then did we scale it to the rest.

ResultTeams use the system instead of bypassing it. An update to a component rolls out to all services, and a new service is built on the existing base.

Multiselect: four component sizes in the Default, Hover, Expanded, Filled and Disabled states
Multiselect: additional use cases — many values, helper text, error
Multiselect: a specification page in Confluence
Design system icons in one grid
Typography: the Open Sans font, a Title scale from 48 to 12 pixels with line heights

My role as Senior Product Designer: the first designer, three teams, process

  • The first designer on the platform: the whole design system and specifications for three development teams.
  • Before the start, I studied the interfaces of six services — I found repeating components and shared patterns.
  • I brought in a second designer and split the component work.
  • I set up a cycle with dedicated front-end developers: design → layout → review with me → test → rollout.
  • I defended the contents of the system at regular demos for team leads and the business.

Result and reflection

Six services on one system that lives without me

All 6 platform services run on the design system. According to feedback from Tricolor managers, operators moved to the unified interface faster than expected. Switching between services stopped being a problem.

BeforeAfter

6 panels built by different teams on Bootstrap

one interface for all 6 services

10 years without a designer and without UX research

dead legacy removed, forms became shorter

features were added on request, without research

a request goes through research and the design system

some elements were unclear or unnecessary to operators

another designer runs the system, and it keeps evolving

DREAMPlatform Scrambler in the platform interface
DREAMPlatform Scrambler
DREPLUS — digital rights management in the platform interface
DREPLUS — digital rights management

What I took from the project

In a project where there was no designer before you, a full system for six services will not be approved on a promise. The audit of all six gave a picture, but approval opened only with a working pilot. First you prove the value on a small scale, then you scale.

A design system is not a component library but a tool for coordinating teams. Without an approval process, the components would have stayed in Figma.