• Work experience

Aalto University (2024-2026)

BDI (2021-2024)

• Get in Touch

Project Summary: Scaling Safe AI in Education

As the lead UX/UI Designer, I led the end-to-end design of Aalto University's accessible, privacy-first AI platform. Collaborating closely with researchers and developers, I owned the user flows, accessibility testing, and UI design to create a secure alternative to public tools like ChatGPT, ensuring sensitive university data remains protected.To drive real-world usage, I designed a specialized Student Course Chat integrated directly into the university's Moodle interface. This feature empowers professors to securely train a localized LLM exclusively on their course materials, providing a safe, closed-loop learning environment that successfully tripled student adoption of our internal AI tool.

The Problem: GDPR Risks and a Breakdown in Trust

When generative AI exploded, University's faced a privacy challenge. Students routinely used proprietary, teacher-created course materials and uploaded them to public LLMs.Through my user interviews and stakeholder surveys, I uncovered three major pain points driving the need for an internal solution:

Role & Strategy: Designing an Intuitive Setup for Educators

Results & Impact: Driving Secure AI Adoption

Next Steps: A Holistic Academic Assistant

Moving forward, we plan to expand the AI chat's capabilities beyond individual courses to serve as a comprehensive academic assistant. The next phase will integrate the secure LLM directly with students' overarching study plans, helping them seamlessly manage weekly tasks, choose future classes, and map out their entire degree journey.

Championing the Aalto Design System to drive adoption, speed, and cost savings.

I joined the Aalto Design System as a hybrid Product Owner and UX Designer to transform a foundational component library into an actively used product standard. My role blended hands-on component design with relentless advocacy, ensuring the system was embraced by internal staff, Aalto startups, and external agencies.To remove friction and make the system truly approachable, I built ready-to-use templates that empowered teams to build on-brand digital touchpoints faster and cheaper. Ultimately, bridging the gap between design and real-world application drove massive adoption across the organization and contributed to the system winning a Red Dot Award in Europe.

The Problem: Stakeholder needs and design system awareness

When I joined the team, the Aalto Design System was roughly 75% complete. While a solid foundation existed, my hands-on collaboration with stakeholders on their active projects revealed critical gaps. The system was missing components (such as custom date pickers and collapsible data cards) and the ready-made templates required for real-world application.This gap between a static library and a usable system created two major bottlenecks:

Role & Strategy: Designing with the Stakeholders

Results & Impact: Unifying the Brand and Driving Adoption

Next Steps: Education and AI Integration

Pushing Boundaries with AI: I am currently testing the limits of the Aalto DS by applying its components to design a new AI interface, exploring how the established system scales to support advanced, conversational UI patterns.

• Branding
• YouTube
• Frammer

Branding for personal youtube project

Elevating 'The Painted Topher' with a cohesive visual design system to drive brand recognition across YouTube, social media, and web.

The mission: breaking away from the ordinary

To support the miniature painting community by removing the stress and intimidation of the hobby. By applying core UX principles to understand painter pain points, the goal was to design an approachable brand and video format that proves painting is accessible to anyone.

The goal was to make a consistent, bold brand in all touchpoints

To cut through a highly saturated YouTube market, I developed a clean, bold, and consistent visual standard. This cohesive design system was applied across all touchpoints. From channel art and social media to in-video edits, graphics, and subtitles, we ensured the educational content was effortless to follow.Established immediate viewer trust and professional credibility, resulting in 1,300 new subscribers and full channel monetization within the first three months.

The Challenge: Standing Out in a Saturated & Intimidating Niche

The Market Saturation: The miniature painting space on YouTube is incredibly crowded. With countless creators fighting for attention, a generic or amateur aesthetic means a channel will instantly get lost in the feed.Many existing tutorials suffer from high cognitive load. Between chaotic thumbnails, inconsistent typography, and cluttered in-video edits, the content often makes the hobby feel stressful and deeply intimidating to beginners.Viewers judge the difficulty of the content by the quality of the design. If the branding looks complicated or messy, potential subscribers will simply scroll past, assuming the painting process will be just as frustrating. The brand needed to visually promise a stress-free, accessible experience before the user even clicked "play."

Designing for Clarity: Building a bold, approachable brand

To cut through the noise of a crowded (and typically dark-mode) YouTube feed, I started by building a high-contrast color system. I grounded the brand in a deep Obsidian Blue, but injected high energy with vibrant pops of Electric Lime and Iris Blue. This wasn't just for aesthetics; it was a deliberate accessibility choice to ensure thumbnails would instantly grab attention as viewers scrolled.

Next came the challenge of scale. YouTube assets live everywhere from tiny mobile feeds to full-screen video edits. I anchored the brand’s typography in Poppins, a modern geometric sans-serif that maintains exceptional legibility whether it's anchoring a channel banner or flashing on-screen as fast-moving subtitles.But a brand is only as good as its execution, and I knew that I needed to focus on painting, not pushing pixels every week. So, I translated these visual choices into a scalable, plug-and-play asset library. Utilizing distinctive, flexible mech silhouettes for the logo system, I built out standardized templates for video covers and social posts.

Next Steps: Scaling & AI Integration

Summary

Aalto students, Staff, and prospective students were drowning in fragmented newsletters split across two platforms mid-migration, with no way to control what they received or when. I led research, design, and testing on a self-serve subscription system that lets users choose their own topics and delivery timing, a problem Aalto had wanted to solve for five years. It is now heading into launch and will be live in the year 2026.

Problem

The trouble started upstream of any design decision: Aalto was mid-migration from Microsoft to Salesforce Marketing Cloud, a necessary, security-driven platform shift, but one that left newsletters scattered across two systems with no single source of truth. Individual schools filled the gap the only way they could; they were building their own manually curated "link list" newsletters, which only fragmented attention further. Some of these got read. Most didn't.What surprised me most was how far people had gone to cope. In interviews, I heard about people forwarding newsletter content to their own browsers just to save it somewhere findable, and others keeping physical notes because there was no digital way to relocate information once it had scrolled past. Staff and teachers, managing even heavier newsletter volume than students, felt it hardest. This wasn't a new complaint; it had been a known, unsolved problem at Aalto for five years.

Role & Strategy

I led the research and design end-to-end, working closely with the team's data analyst, IT, product owners, the team lead, and developers. Research spanned fifteen student interviews, five staff interviews, and five teacher interviews, backed by click-through data pulled directly with IT to see not just what people said, but what they actually opened. Rather than trying to solve everything at once, we scoped the project in two phases: students and staff first, prospective students later.

Approach

One insight shaped almost every downstream decision: people didn't just want relevant content; they wanted it on their own terms. Several interviewees specifically mentioned wanting their newsletters to land on Friday mornings or over the weekend, frustrated that updates kept arriving at the wrong moment, buried under everything else in their inbox. That reframed the problem from "what should we send" to "when do people actually want to receive it," and became the core design principle behind the system.I designed a subscription center around a simple side-navigation interface. Users can build up to six subscriptions, each pulled from Aalto's existing tags, so there was no new taxonomy to learn or maintain. They choose their own delivery timing per subscription and receive updates in two ways: a personalized email through Salesforce Marketing Cloud, and an in-context notification, which is a small indicator in the top right of Aalto's main site. The design was prototyped and retested with the same people who'd been interviewed, closing the loop between what they'd asked for and what they were actually getting.

Challenges

Two points of friction shaped the project in ways I hadn't anticipated. Early on, we needed real alignment with developers on exactly how the Aalto tags should function within the system. This conversation took longer than expected and delayed initial progress until we got it right. Separately, school communications staff pushed back in workshops, not out of resistance, but because they wanted more control over their own tagging rather than being limited to a fixed set. That feedback didn't get shelved; it became a second-phase feature, letting individual schools define and push their own custom tags.

Results & Next Steps

Usability testing surfaced something I hadn't expected: people were conflating "subscription" with "newsletter" itself, treating the two as interchangeable even though the system worked differently under the hood. I reframed the interface's language until that distinction was unambiguous. Beyond the interface fixes, the workshops themselves told their own story; stakeholders weren't just tolerating the idea, they were genuinely eager to adopt it. The project is heading into launch now, and I'll be updating this case study with real adoption data once it's live.