Hi, I’m Lylian Miller, UX Engineer and Product Designer.
I lead product experiences across enterprise software, SaaS platforms, and AI-powered tools, with deep expertise in UX strategy, scalable design systems, and complex workflow design. I work at the intersection of product design and engineering, and my visual design background strengthens the clarity, quality, and impact of every interface I build.
Case Studies
Survey Studio
AI-native platform covering every stage of the survey lifecycle, from authoring and screener logic to in-survey experiences and post-survey workflows.
role
Scope
Stack
Lead Product Designer
UX, Design System, Front-end architecture
Figma, Design Tokens, LLM Capability Research
context & Challenge
A platform built before it was officially a project
Survey Studio didn’t begin as a sanctioned initiative. Before leadership greenlighted the project, I had already invested time in a proof of concept, mapping the information architecture, identifying the critical workflow gaps across survey authoring, screener logic, and post-survey stages, and validating design directions in Figma. By the time the team received a formal go-ahead, I had already eliminated weeks of ramp-up.
The core challenge wasn’t just design. It was orchestrating a connected system across multiple internal workflows, each with its own stakeholder expectations, technical constraints, and research contexts while keeping the experience coherent and scalable for a non-technical user base of qualitative and quantitative researchers.
Design leadership
Component-first, system-minded
Every component was designed from scratch in Figma, with production use in mind from day one. Each pattern was built to extend across the full platform and into adjacent applications that share the same design language.
Brand & Identity
Logo and visual identity
I created the Survey Studio logo myself, drawing on 17 years of cross-departmental experience spanning software engineering and marketing. Identity wasn’t an afterthought; it was part of building something that felt like a real product, not an internal tool with a temp name.
AI-Accelerated Workflow
While building the design system, I worked in parallel using AI tooling, specifically Claude Design to handle CSS generation and front-end specification in the background. This let me focus simultaneously on UX flows, new screens, and engineering decisions without bottlenecking on documentation. The result was a faster, higher-coverage handoff without sacrificing design precision or developer trust. A working method I’ve refined to stay two steps ahead on complex multi-track projects.
UX Engineering
Thinking in front-end, not just in frames
My contributions extended beyond design deliverables. I shaped how components would be built, accounting for state management, branching logic across multi-step flows, and the performance implications of reusable pattern libraries. When front-end decisions needed to be made, I made them informed by both design intent and engineering tradeoffs, reducing back-and-forth between disciplines.
Survey Studio was also the first application in the suite where I introduced LLM-assisted features into the product itself, not just the workflow. These capabilities accelerated researcher throughput while preserving the accuracy and trust standards required by our researchers and clients.
SurveyGPT
An AI-powered survey design application built to make advanced LLM capabilities feel intuitive, trustworthy, and immediately useful for enterprise research teams.
role
Scope
Stack
Lead Product Designer
End-to-end: UX, UI architecture, component systems
Figma, Design Tokens, LLM Capability Research
context & Challenge
Designing for AI before there was a playbook
SurveyGPT arrived at a moment when most enterprise software companies were still trying to understand what AI-powered products should even feel like. There were no established UX conventions for LLM-driven workflows. I had to build the design language from the ground up, which meant investing deeply in understanding how these models actually worked, what they could and couldn’t do reliably, and how to communicate that honestly to users who had never interacted with generative AI in a professional tool.
The challenge wasn’t just designing the UI. It was establishing trust between the user and the system, making automation feel like a collaborator rather than a black box, and building a product that could scale as the underlying technology continued to evolve rapidly underneath it.
Approach
Automation balanced with human control
Design focus
Clarity, trust, and long-term scalability
Collaboration
Direct engineering partnership for handoff quality
Staying Ahead of the Curve
When SurveyGPT was in development, most design teams were still treating AI as a novelty. I treated it as a discipline. I invested significant time researching LLM capabilities, behavioral patterns, failure modes, and UX conventions emerging from the broader AI product ecosystem. That foundation let me design interactions that aged well, scaling with new model capabilities rather than becoming obsolete by them. My interfaces were built to flex, not to lock in assumptions about what the AI could do at a single point in time.
Engineering Collaboration
Front-end fluency
My background in front-end development shaped every deliverable I handed off. I understood how components would be structured in code, what edge cases developers would run into, and where design decisions could create unnecessary engineering complexity. That awareness made the collaboration tighter and the handoffs faster, reducing the revision cycles that tend to erode timelines on AI-heavy products where the requirements are already moving targets.
Accessibility and usability weren’t layered on at the end. They were built into the component logic from the start, ensuring that the AI-driven interactions met the standards required by enterprise clients while remaining genuinely usable under real working conditions.
AIM
An AI and LLM-powered research tool suite built for external enterprise clients, where the experience directly reflects on the company’s credibility.
role
Scope
Stack
Lead Product Designer
Figma, Design Tokens, Component Handoff
Figma, Design Tokens, Component Handoff
context & Challenge
Enterprise AI, for external clients
AIM is a suite of interconnected research tools, each serving distinct AI-driven functions while sharing a unified design language. Unlike internal products, this was client-facing, where polish and trust aren’t nice-to-haves. I led design across the full suite, simplifying sophisticated research workflows into experiences that felt credible and efficient without flattening the depth each tool required.
UX Strategy
Patterns built for a moving target
With AI capabilities evolving rapidly through the project, I designed scalable patterns flexible enough to absorb new features without structural redesigns, and applied reusable system thinking across every tool in the suite to keep the experience coherent as it grew.
Engineering Collaboration
Production-ready from the start
My front-end background shaped every handoff. I anticipated implementation complexity early, framed design decisions in terms engineers could act on immediately, and kept the gap between what was designed and what shipped as tight as possible.
Capability Identity
I designed the logo for one of AIM’s core capabilities, an extension of my cross-disciplinary background across software engineering and marketing. When a designer can own the full surface area of a product, from UX architecture to the mark that represents it, the result is a more coherent product and a stronger design voice at every stage.
Engineering Collaboration
Production-ready from the start
Close partnership with engineering was central to how I worked on AIM. My front-end background shaped how I framed design decisions, anticipating implementation complexity early and keeping handoffs tight. The goal was always to reduce the distance between what was designed and what shipped.
icanmakeitbetter
A multi-tool platform supporting long-term community engagement across qualitative and quantitative research, serving both enterprise clients and the participants they depend on.
role
Scope
Stack
Lead Product Designer
Product design, marketing design, motion graphics
Figma, Motion Design, Marketing Design, Adobe Suite
context & Challenge
Two audiences, one platform, no shortcuts
icanmakeitbetter served a genuinely split audience: enterprise clients managing long-term research programs, and the community participants those programs relied on to stay engaged over time. Designing for both meant holding two very different sets of needs in tension simultaneously. The client-side demanded the efficiency, hierarchy, and data density of a professional research tool. The participant-side required warmth, clarity, and enough motivation to keep people coming back across weeks or months, not just a single session.
Product Design
Consistency across a growing ecosystem
I led design across a broad set of qualitative and quantitative research tools, applying reusable component thinking to keep the platform coherent as its feature set expanded. Scalable UX patterns and a shared design language meant each new tool could ship without fracturing the experience users had already built familiarity with.
Engineering Partnership
Production-ready at every stage
Close collaboration with engineering kept design decisions grounded in what could actually ship cleanly. My front-end background helped translate design intent into terms the team could act on directly, tightening the feedback loop and keeping the gap between design and implementation narrow.
Beyond Product: Marketing & Motion
My scope extended well past the product interface. I contributed marketing design work and created motion graphics aimed at sustaining community engagement over the long term, a non-trivial challenge when participant retention directly affects research quality. This is where 17 years of cross-departmental experience across product, engineering, and marketing converge into something most designers can’t offer: the ability to hold the full brand and user experience in view at the same time, and design for both without losing coherence in either direction.
QualConnect
A video-enabled qualitative research application extending the icanmakeitbetter ecosystem, designed to support richer, more complex research interactions without adding friction for either clients or participants.
role
Scope
Stack
Lead Product Designer
UX, information architecture, branding, front-end collaboration
Figma, Design Tokens, Component Handoff
context & Challenge
Richer research, without added complexity
QualConnect introduced video capabilities into a platform that participants and clients already had expectations around. Adding a medium as demanding as video to a qualitative research workflow creates real UX risk: it raises the cognitive load, introduces new failure states, and can erode the trust that longitudinal research communities depend on. The challenge was making video feel like a natural extension of the experience rather than a bolt-on feature, while keeping the interaction patterns consistent with the broader icanmakeitbetter ecosystem users were already familiar with.
UX & Information Architecture
Structure that scales with complexity
Video-based research workflows carry more moving parts than text-based ones: session states, recording status, participant readiness, async review flows. I refined the information architecture to keep those layers legible without surfacing everything at once, letting users navigate complexity at their own pace without losing orientation.
Branding & System Cohesion
A distinct identity within a shared system
QualConnect needed its own brand presence while remaining recognizably part of the icanmakeitbetter family. I handled the branding work alongside the product design, ensuring that the visual identity reinforced the platform’s positioning as a premium qualitative capability without breaking the design consistency users expected across the ecosystem.
Front-end Collaboration
As with the broader platform, my front-end background shaped how I delivered design for engineering. I thought through component states, edge cases, and implementation tradeoffs before handoff rather than after, which kept integration work moving cleanly even as the video feature set introduced more technical surface area to account for.
Insight Ops Suite
An internal operational hub designed to support the Insights Ops team across every phase of quantitative and qualitative project management, from intake to delivery.
role
Scope
Stack
Lead Product Designer
UX strategy, information architecture, low to high fidelity
Figma, Design Tokens, Component Handoff
context & Challenge
Operational complexity that needed a foundation
Internal operational tools accumulate complexity over time. Insight Ops Suite served a team managing both quantitative and qualitative research projects simultaneously, each with distinct workflows, handoff points, and stakeholder requirements. The challenge wasn’t cosmetic. It was structural: the underlying task logic needed to be untangled before any interface decisions could be made responsibly, or the design would simply inherit the same friction it was meant to resolve.
Process
Before any high-fidelity work began, I led a dedicated UX phase grounded in low-fidelity information architecture and user flows. This step isn’t always given the space it deserves, but for an operational suite with layered task logic it was essential. Working at low fidelity forced clarity on what the workflows actually needed to do, surfaced decision points that would have become expensive problems in production, and gave stakeholders a shared model of the system to align around before a single component was designed.
UX & Interface Design
Simplifying without oversimplifying
Operational tools fail when they try to hide complexity from users who actually need it. I focused on improving interface structure and hierarchy to make dense workflows navigable at a glance, defining scalable UX patterns and reusable design solutions that could grow with the suite without requiring redesigns as new project types were added.
Engineering Collaboration
System thinking through to delivery
Design decisions were made with implementation in mind throughout. Close partnership with engineering meant that the scalable system thinking built into the UX phase translated into production-ready solutions, keeping the design coherent from the first wireframe through to the shipped product.
Sentry
A respondent gateway experience managing privacy communication and automated quality controls at the point of entry, before respondents reach a single survey question.
role
Scope
Stack
Lead Product Designer
UX strategy, low-fidelity flow mapping, accessible UI design
Figma, Design Tokens, Component Handoff
context & Challenge
Where trust is earned before the experience begins
Sentry sits at one of the most consequential moments in any research workflow. A gateway experience has to communicate privacy policy clearly, execute automated quality checks without surfacing unnecessary friction, and do all of it quickly enough that the respondent never feels interrogated. Get any of those wrong and data quality suffers before a single question is answered. The design challenge was balancing protection for the platform with a seamless, trust-affirming path forward for the respondent.
Process: Logic Before Layout
As with Insight Ops Suite, I started in low fidelity, mapping information architecture and user flows before any visual design work began. For a gateway experience driven by conditional logic and policy requirements, this phase was especially important. The branching paths through privacy communication and quality control checks needed to be airtight structurally before they could be made seamless experientially. Low-fidelity work exposed edge cases and decision points early, when they were still inexpensive to resolve.
Trust & Privacy Design
Transparency without friction
Privacy communication is one of the hardest things to design well. Most interfaces either bury it in legalese or disrupt the experience so aggressively that respondents abandon the flow entirely. I focused on clarity and pacing, designing interfaces that communicated policy honestly while keeping the path forward visible and low-effort at every step.
Accessibility & Precision
Designed for everyone who passes through
A respondent gateway has a wide and uncontrolled audience. Accessibility wasn’t optional. I applied accessible, user-centered interface standards throughout, ensuring that the experience held up across device types, assistive technologies, and varying levels of digital familiarity, without compromising the logic-driven precision the quality control layer required.
MySoapBox
A consumer-facing survey platform built around respondent participation and reward-based engagement, designed for everyday users with no assumed technical familiarity.
role
Scope
Stack
Lead Product Designer
UX, UI, motion graphics, marketing creatives, SEO
Figma, After Effects, Illustrator, Photoshop
context & Challenge
Designing for the widest possible room
MySoapBox is the consumer end of the research ecosystem, where the audience is not researchers, operators, or enterprise clients, but everyday people choosing to spend their time in exchange for rewards. That shift in audience changes everything about how design decisions get made. Complexity that an internal power user tolerates becomes abandonment for a consumer. Friction that goes unnoticed in a professional context directly erodes participation rates when the user has no obligation to stay.
UX & UI Design
Simple on the surface, solid underneath
I focused on translating complex platform requirements into experiences that felt approachable and easy to navigate without oversimplifying what the product needed to do. Scalable UX patterns and reusable components kept the interface consistent as the platform grew, while structured interaction design ensured the experience remained predictable for returning users.
Engagement & Retention
Keeping participation voluntary and motivated
Reward-based platforms live and die by whether users feel the exchange is worth it. Beyond the visual design, I considered how interface choices affected perceived value and momentum through the experience, reducing friction at the moments most likely to cause drop-off and reinforcing the sense of progress that keeps voluntary participation going.
Beyond the Product: Visual, Motion & Marketing
My scope on MySoapBox extended well past the product interface. I created motion graphics and visual design work aimed at driving community engagement, and produced marketing creatives for Facebook ad campaigns and SEO initiatives. This kind of full-funnel design ownership, from the moment someone sees an ad to the moment they complete a survey, is only possible when product, visual, and marketing disciplines are held by the same person. It keeps the brand voice consistent and removes the gaps that tend to form when those responsibilities are split across teams.
SimplerDAO
A decentralized autonomous organization interface designed to make Web3 governance concepts feel accessible, trustworthy, and usable without leaning on hype or unnecessary complexity.
role
Scope
Stack
Product Designer, Branding
UI design, interaction patterns, systems thinking, branding
Figma, Design Tokens, Web3 Domain Research
context & Challenge
A space defined by friction, ambiguity, and misplaced confidence
Web3 product design occupies a unique position in the UX landscape. The technology is genuinely complex, the user base ranges from crypto-native to completely unfamiliar, and the ecosystem has a long history of interfaces that obscure rather than clarify. DAOs in particular require users to understand governance mechanisms, voting logic, and decentralized ownership models before they can meaningfully participate. The design challenge was making all of that legible without flattening it, and without defaulting to the aesthetic shortcuts that give Web3 products their reputation for being style over substance.
Domain Knowledge
Designing from the inside out
My background in crypto meant I wasn’t learning the domain while designing for it. I came into the project with a working understanding of how DAOs function, what decisions governance participants actually need to make, and where decentralized systems tend to break down for non-technical users. That context shaped every interaction pattern and information hierarchy decision from the start.
Branding & Identity
Trust built into the visual language
In a space where trust is hard-won and easily lost, visual identity carries more weight than it does in established product categories. I contributed to the branding alongside the product design work, ensuring that the interface communicated credibility and clarity from the first impression, not just in the flows themselves but in the aesthetic language that framed them.
Emerging Technology, Grounded UX
What SimplerDAO represents in this portfolio is range. Every other project here is enterprise or consumer research. This one is Web3 governance, a completely different problem space with a different user mental model, different trust dynamics, and a different tolerance for ambiguity. Designing well across contexts like these requires more than a transferable process — it requires genuine curiosity about how new technologies change the relationship between a product and the people using it.
Fleet Writers Room
A fully designed and developed website for an Office of Naval Research supported program exploring next-generation defense concepts for Navy-related audiences, owned end-to-end from concept to delivery.
role
Scope
Stack
Designer, Developer
UX strategy, visual design, content hierarchy, front-end development
Figma, HTML, CSS, JavaScript
context & Challenge
Credibility is the product
Fleet Writers Room required a digital presence capable of representing a serious, government-adjacent research program to a specialized Navy audience. In this context, visual polish and structural clarity weren’t aesthetic preferences, they were credibility signals. A poorly organized or visually undisciplined site would undermine the program’s authority before a visitor read a single word. The design had to match the weight of the subject matter while remaining genuinely easy to navigate for an audience that values precision over decoration.
End-to-end Ownership
This project had no handoff. I carried the work from initial concept through UX structure, visual design, content hierarchy, and into production. Full ownership at this level is where the UX Engineer value proposition is most visible: when the same person who defined the information architecture also wrote the code that implemented it, nothing is lost in translation. What was designed is what shipped, with exactly the intent it was built with.
UX & Visual Design
Structure that earns attention
I established a clear content hierarchy that guided visitors through complex program information without overwhelming them. Visual design decisions were made to reinforce credibility and professionalism at every level, from typographic choices to layout rhythm, ensuring the experience felt considered and authoritative rather than templated.
End-to-end Website Design
Designed and built by the same hands
Moving from Figma into production without an engineering team required disciplined decision-making throughout the design phase. Every layout choice, interaction pattern, and visual detail was made with implementation in mind from the start, resulting in clean, production-ready front-end code delivered independently and on scope.
Design Systems
Systems, flows, and design operations built across a decade and a half of practice — from atomic foundations to AI-accelerated delivery, grounded in experience rather than replaced by it.
role
Scope
Stack
Cross-project, all products
Atomic design, design tokens, AI, low-fidelity flows
Figma, Claude Design, Google Stitch, Subframe
Foundation
Systems that outlast the sprint they were built in
Across every product in this portfolio, design systems work has been a constant thread. From small-scale component libraries to complex token-based foundations spanning multiple interconnected applications, I build systems with the same principle in mind: the decisions made in the system layer should make every individual design decision faster, more consistent, and easier to defend. Atomic design methodology gives that structure a rigorous foundation, while reusable component logic and token-based architecture ensure the system can scale without fracturing.
Fidelity & Strategy
The right resolution for the right moment
Not every project warrants the same approach, and knowing which strategy to apply, and when. Is a skill that only comes with experience. For complex operational products with ambiguous requirements, I start in low fidelity: flow maps, IA diagrams, and wireframes that align teams and surface structural problems before a single component is designed. For POCs and time-sensitive initiatives, I move directly into high fidelity to demonstrate value quickly and give stakeholders something concrete to react to. When a company is early-stage or validating direction, the deliverable is momentum, not polish. Reading the project context and calibrating accordingly is as important as the design execution itself.
Front-end Collaboration
Systems built for the people who implement them
A design system is only as strong as its adoption in code. My front-end background shapes how I document components, define token structures, and communicate implementation intent — making the gap between design and engineering output as narrow as possible. Handoffs read like instructions that developers can act on, not approximations they have to interpret.
AI Tools in the Design Workflow
I actively explore and adopt emerging AI tools that change what design systems work can look like in practice. Claude Design, Google Stitch, and Subframe each represent a different point in that landscape. From CSS and front-end specification generation to component scaffolding and rapid UI exploration. I use these tools where they create genuine leverage: handling spec output, accelerating token documentation, and pressure-testing design decisions against production constraints faster than manual methods allow.
Collaboration-forward Built with Researchers.
Design Work
Designed, engineered, and shipped end to end.
Websites built from brief to live URL, across enterprise and consumer work. Design decisions made with implementation in mind. Front-end written with design intent preserved. No translation loss between disciplines.
Tools
The right tool for the brief. Static sites on Cloudflare Pages for speed and ownership. Webflow and WordPress when a team needs a CMS. Figma for design, code when the implementation demands it.
Drawn by hand, iterated with AI, built to engage.
Social creatives, campaign assets, and motion-led visuals designed to capture attention quickly and support acquisition, retention, and brand trust across B2B and B2C audiences.
process
Some pieces begin with illustration. Others start with AI-assisted exploration and are then shaped, edited, and refined by hand. The skill is in choosing the right approach for the brief and turning it into work that feels distinctive, purposeful, and ready to perform.
Brand vision, made consistent across every medium.
Branding work that turns a clear vision into a practical visual system. Designed to stay consistent across product, marketing, print, and digital touchpoints.
Identity
Logos, marks, and brand systems designed to scale across every application, from favicons and social assets to presentations, packaging, and large-format placements without losing clarity, recognition, or intent.
Interaction showcase
Product experiences in motion
Animated work developed through client projects, spanning interface interactions, product storytelling, and supporting brand moments.
Visual explorations
llustration and painting projects that complement my product design practice, and occasionally support consumer-facing concepts.
