Elizabeth Kaznadey | Product Designer

Current Role

Designing products at Fliplet

For 4.5 years I've been the product designer at Fliplet, a low-code/no-code platform where people build apps using templates and components without needing developers. Our clients range from enterprise (mostly law firms) to self-service users building their own solutions.

I was the only designer on the product side, so I owned the process end to end: research, ideation, prototyping, testing, and working through implementation with engineers. I partnered closely with the product owner, CEO, engineering, growth, and customer success. Design decisions didn't happen in isolation.

The work covered a lot of ground: new features, design systems, templates with complex logic, and improvements to existing product areas. Alongside building new things, I continuously reviewed how the product was performing, watching sessions, surfacing problems, and collaborating with growth on experiments.

Over the last year, my process shifted. I started using AI tools heavily for research, prototyping, and even building functional features myself. It's made me faster and more technical than I've ever been.

I often pushed for changes before they were on the roadmap. I advocated for a platform redesign for years. It wasn't a priority until recently, but we're now implementing it with new features. When I saw a better approach, I made the case for it.

Alongside larger projects, I've shipped dozens of smaller improvements: UI refinements, flow fixes, experiments with the growth team. Not everything needs a case study, but it adds up.

How I Work

My Approach

I believe good design comes from looking at things from different angles. Testing ideas, trying alternatives, staying open.

User-Centered

Users will always show you the truth. If they don't understand how something works, it's not their fault, it's bad design.

Collaborative

I do my best work with a team. Not because I can't work alone, but because ideas get better when you think out loud together. You speak, you sketch, you react, and suddenly you see things you'd miss on your own. Reviewing work in isolation makes it hard to spot the problems.

Fast Iterations

I like fast iterations, clear goals, and the autonomy to figure out how to get there. I don't work well when I'm just told what to do. I want to be a partner, not a pair of hands.

AI-Enhanced

I use AI tools heavily for research, prototyping, and building functional features. It's made me faster and more technical than ever.

Projects

Fliplet Rebuilt

View Prototype

Full platform redesign driven by AI.

Click to open prototype

Context

Fliplet wanted to shift to an AI-driven platform, similar to Lovable or Bolt. The goal: let users build apps through conversation instead of drag-and-drop components and screen templates.

What I did

I'm leading the UI/UX for this project. That means extensive competitor research, designing the core experience (chat interface, response patterns, how users provide structured data, apply or roll back changes) and building working prototypes.

The interface is conversational. Users describe what they want, select elements on screen to refine them, and let AI handle the building. AI can suggest improvements, generate sample data, and test the app. There's an appearance settings panel for global styling. Users can even extract a visual style from an uploaded image.

Every decision has constraints: what's technically possible, what's intuitive, what's reliable enough for enterprise. And not all clients are ready for AI. The experience needs to work for people at very different comfort levels.

Outcome

First phase is launching internally soon. Client-facing launch is planned for mid to late spring. I'm actively working on it now. It's moving fast.

Design System

View Prototype

Building a flexible system that works for humans and AI.

Click to open prototype

Context

Fliplet's platform had grown over many years, with features added by different people at different times. There was no unifying system, and inconsistencies had piled up. It was never the priority to fix, but it showed.

What I did

I audited existing patterns and designed a system that brings consistency without making everything feel unfamiliar. But I didn't just build a Figma library. I built something that could be read by AI.

The system consists of design tokens as JSON, individual JSON files for each component, global CSS and JavaScript, a visual documentation web app, markdown implementation guides, and a Claude.md file with AI maintenance instructions.

I use it to build new features myself via MCP, connecting to Fliplet and producing production UI directly. It's also the foundation for Studio V3, and I'm training AI to produce components using these patterns.

To capture the existing system, I built a tool that takes screenshots, extracts code via Magic Patterns, and converts it into structured design components.

Outcome

It's ongoing. I keep finding new patterns and challenges. But right now, I can produce any layout in minutes, fully on-brand and consistent.

Action Builder

View Prototype

A tool for creating app logic, now powered by AI.

Click to open prototype

Context

Action Builder lets users create actions for their apps, things like "when this button is clicked, send an email." I started working on this a few years ago with a visual approach: blocks where users connect data, link screens, set dynamic options.

Then AI appeared, and we realised users don't need to build actions manually anymore.

What I did

The project shifted from a visual block-based builder to a conversational interface. Users now see a list of their actions and can ask AI to create new ones or modify existing ones. They still guide the AI (providing data, specifying connections) but the building happens through chat.

I built the feature myself using the embedded app approach. Engineers provided APIs and documentation, I connected to Fliplet via MCP, used my design system, and created the working UI. Then I passed it to engineers to wire up the backend.

Partway through development, I made the call to split the feature into two screens. The code was partially built, but the UX wasn't right. I used AI to replicate and restructure the work, and we shipped the better version.

Outcome

A project that completely changed direction midway. That shift forced me to rethink assumptions about how users interact with complex logic. When AI handles the complexity, the interface can become much simpler.

Publishing Flow

Rebuilding how users publish their apps to the App Store, Google Play, and web.

Context

Users kept running into failed builds and app store rejections, often only discovering issues after they'd already started the process. It created frustration and a constant stream of support requests.

What I did

I redesigned the entire publishing experience from scratch. The new flow validates each step before users move forward, so problems surface early, not after a build fails. We shifted the model: instead of trying to control the app store submission process (which changes constantly), we now build the app and give users clear guidance to publish it themselves. We also automated manual headaches, like certificate generation that now renews automatically.

The original plan was to keep the existing structure with separate tabs for each channel. I pushed for a dashboard view instead, giving users a full picture of all their publishing in one place. It became the main screen.

This was a complex system: three publishing channels, each with multiple states, and users needed to complete all three in one flow without getting lost. A lot of the work was figuring out what level of automation engineering could support, and how to make a technically dense experience feel approachable.

Outcome

Reduced number of failed builds and support requests. Users now see issues clearly at each step before they become problems.

Publishing Dashboard

iOS Dashboard

iOS Steps

Complete Steps

Submission History

App Templates

Fully working apps that customers can customise without breaking functionality. I built these end-to-end: research, flows, logic, UI, testing, and launch.

Context

App templates are one of Fliplet's core offerings. These aren't starting points or wireframes. They're fully working apps that customers can modify without breaking functionality. Each template typically has 40-50 screens, multiple user roles (admin, user, manager), and needs to cover every state and scenario.

I've been working on templates for years. It's where a lot of my product thinking was shaped.

What I did

I built templates end-to-end: research, competitor analysis, user flows, documentation, stakeholder alignment, building in Fliplet Studio, styling, custom code where needed, testing, and launch.

The challenges

The hard part is holding two things at once: the template needs to be complex enough to be a real, working app, but simple enough that customers can adjust it without breaking anything.

UI is limited by component settings. You can't add custom styling if it means users would need code to change it later. So you're designing within constraints, making something that works and looks good with what's available.

With AI, the speed has increased. We also tried a new approach: working live on calls with developers, applying changes together instead of reviewing after. It keeps momentum.

Outcome

Templates are used by enterprise and self-service clients to launch apps faster. This work shaped how I think about flexibility, constraints, and designing for real use, not just demos.

Template examples All Templates

Capacity Management

Tracking employee availability and location, integrated with the directory solution.

Learning Solution

An online training tool where companies upload courses and employees complete them to earn certificates. Complex logic around course building, correct answers, scoring, and completion tracking.

Task Management

A tool for creating and assigning tasks, tracking progress, with public and private lists, templates, and a dashboard.

Event Solution

A full event management app with attendees, speakers, exhibitors, and admins, five different user roles with different permissions. QR check-in, sessions, agendas, gamification, and data exports.

Emergency Management

Built for law firms to prepare for crises. Step-by-step flows, assigned tasks, notifications, document uploads, chat, and a lawyer-facing dashboard to track progress.

About Me

A bit more about me

I'm originally from Kyiv, a city full of creative and open-hearted people. I moved to London almost two years ago, and I'm still building my life here: meeting new people, exploring the city, finding my spots.

Outside of work, I love being outdoors. Parks, long walks, new places. Travel recharges me more than anything. I run now too. You can't not start running in this country. I draw sometimes, and spend a probably unreasonable amount of time growing vegetables from Ukrainian seeds on my terrace.

I recently got a puppy with my partner, which has taught me a lot about consistency, patience, and very early mornings.

I care about people. I think everyone is a unique mix, and I value real relationships over surface-level ones.

Say Hi

Let's work together

Looking for a great team where good work happens together.
I'm open to new opportunities and always happy to chat.

Email elizaveta.kaznadey@gmail.com Copied!

Phone +44 778 073 7462

LinkedIn /in/elizabeth-kaznadey

CV View | Download

Location London, UK

my name is Elizabeth

Designing products at Fliplet