I'm Elina — Senior Product Designer & UX Researcher in Germany.
Ich bin Elina — Senior Product Designer & UX Researcherin in Deutschland.
M.Sc. in Design & Usability — a top-1000 world university, QS ↗
B.Sc. in Applied Informatics, Data Science — top 300, QS regional ranking ↗.
M.Sc. in Design & Usability — Top-1000-Universität weltweit, QS ↗
B.Sc. in Angewandter Informatik, Data Science — Top 300 im regionalen QS-Ranking ↗.
12 case studies, one collection — group it by problem, industry, business model, platform or skill; projects live in several groups at once.12 Case Studies, eine Sammlung — gruppiert nach Problem, Branche, Geschäftsmodell, Plattform oder Skill; Projekte gehören mehreren Gruppen an.
Material 3, fully: Calendar, Emails, Tasks (enterprise suite).
HIG, fully: Verbly, blogi. Platform conventions first — flexed only with a reason.Material 3, vollständig: Kalender, E-Mails, Aufgaben (Enterprise-Suite).
HIG, vollständig: Verbly, blogi. Konventionen zuerst — Abweichung nur mit Grund.




Pricing, checkout, payment success and failure, billing, cancel, win-back — every unhappy path designed to the same fidelity as the happy one.Pricing, Checkout, Zahlung erfolgreich und fehlgeschlagen, Billing, Kündigung, Win-back — jeder Unhappy Path in derselben Qualität wie der Happy Path.




Helping people find the thing they want to buy: catalog navigation proven by controlled experiments (n=342+), search that parses real query intent, and product cards that earn trust for high-stakes purchases.Menschen finden lassen, was sie kaufen wollen: Katalog-Navigation, belegt durch kontrollierte Experimente (n=342+), Suche, die echte Query-Intention versteht, und Produktkarten, die bei teuren Käufen Vertrauen schaffen.
AI features I designed inside these products: AI dialogues with live corrections and a tutor mode in Verbly · job matching that explains every recommendation in CareerSide · AI content generation from idea to finished carousel in blogi. How I work with AI myself — AI workflow ↓AI-Funktionen, die ich in diesen Produkten designt habe: AI-Dialoge mit Live-Korrekturen und Tutor-Modus in Verbly · Job-Matching, das jede Empfehlung erklärt, in CareerSide · AI-Content-Generierung von der Idee bis zum fertigen Carousel in blogi. Wie ich selbst mit AI arbeite — AI-Workflow ↓



Churn lives in the unhappy paths, not on the dashboard. In the field it paid off: +25% uptime in low-connectivity areas, −30% missed SLAs (courier app, team result).Churn wohnt in den Unhappy Paths, nicht auf dem Dashboard. Im Feld zahlte es sich aus: +25% Uptime in Funklöchern, −30% verpasste SLAs (Kurier-App, Teamergebnis).




Retention mechanics designed as habit loops: streaks with repair paths, leagues, quiz games and visible progress — motivation without shame patterns.Retention-Mechaniken als Habit Loops: Streaks mit Reparatur-Pfaden, Ligen, Quiz-Spiele und sichtbarer Fortschritt — Motivation ohne Shame-Patterns.




Working dashboards, not vanity charts: geo-analytics where the map takes 70–80% of the screen, role-specific modular widgets, a job-search cockpit and a coach's client workspace.Arbeits-Dashboards statt Deko-Charts: Geo-Analytics mit 70–80% Karte, rollenspezifische modulare Widgets, ein Jobsuche-Cockpit und der Klienten-Arbeitsbereich für Coaches.




The cheapest test for the riskiest assumption first: interviews, surveys n=241/252, concept split-tests, CustDev with 12 C-levels — 800+ participants across studies.Der günstigste Test für die riskanteste Annahme zuerst: Interviews, Umfragen n=241/252, Konzept-Splittests, CustDev mit 12 C-Levels — 800+ Teilnehmende über alle Studien.



Tokens and 50+ documented components — cutting UI implementation time by 40% and keeping every screen consistent across platforms.Tokens und 50+ dokumentierte Komponenten — UI-Implementierungszeit −40% und jeder Screen plattformübergreifend konsistent.



2.5 years on a mobile suite used by large companies: 30+ features end-to-end, +20% user retention two years post-launch.2,5 Jahre an einer Mobile-Suite für Großunternehmen: 30+ Features End-to-End, +20% User Retention zwei Jahre nach Launch.



AI speaking practice with retention that respects adults: three correction modes, streak-repair instead of shame — sessions grew from 8 to 19 minutes.AI-Sprechpraxis mit Retention, die Erwachsene respektiert: drei Korrekturmodi, Streak-Reparatur statt Scham — Sessions wuchsen von 8 auf 19 Minuten.



Habit loops and a trainer-client platform — progress you can feel, not just log. Plus pro-bono WCAG-AA work for a Berlin mental-health nonprofit.Habit Loops und eine Trainer-Klienten-Plattform — spürbarer Fortschritt. Plus Pro-bono-WCAG-AA-Arbeit für eine Berliner Mental-Health-Organisation.
Field-grade maps: offline tiles, geofencing, in-map scanning — plus map-first analytics for two user roles. Researched through ride-alongs with couriers.Karten für den Feldeinsatz: Offline-Kacheln, Geofencing, Scan in der Karte — plus Map-first-Analytics für zwei Rollen. Erforscht durch Mitfahrten mit Kurieren.



Trust-first listings and frictionless checkout — commerce design for products people can't hold before buying.Trust-first-Listings und reibungsloser Checkout — Commerce-Design für Produkte, die man vor dem Kauf nicht in der Hand hält.
An AI agent that explains every job match — and an AI co-writer that keeps the creator the author.Ein AI-Agent, der jedes Job-Match erklärt — und eine AI-Co-Autorin, bei der der Creator Autor bleibt.
Enterprise productivity, fleet logistics and GIS for companies — research-heavy, spec-heavy, design QA to 95% compliance.Enterprise-Produktivität, Flottenlogistik und GIS für Unternehmen — research- und spezifikationsstark, Design-QA bis 95% Compliance.




Consumer products that earn daily use — subscriptions, habit mechanics, honest paywalls and checkout without friction.Consumer-Produkte für tägliche Nutzung — Abos, Habit-Mechaniken, ehrliche Paywalls und Checkout ohne Reibung.




The full subscription loop, B2B and B2C: trial, paywall after first value, billing, limits that never interrupt mid-task.Der volle Abo-Loop, B2B und B2C: Trial, Paywall nach erstem Mehrwert, Billing, Limits, die nie mitten im Task unterbrechen.




Native iOS & Android to HIG and Material — 30+ shipped features, 100+ spec'd screens with all states and edge cases.Natives iOS & Android nach HIG und Material — 30+ Features, 100+ spezifizierte Screens mit allen States und Edge Cases.











Responsive platforms and dashboards with the same discipline as mobile: evidence, design systems, edge cases.Responsive Plattformen und Dashboards mit derselben Disziplin wie mobil: Evidenz, Designsysteme, Edge Cases.




Two CustDev cycles (12 C-level + external directors) shaped an enterprise product pre-MVP — and helped raise retention +20% two years post-launch.Zwei CustDev-Zyklen (12 C-Level + externe Direktoren) prägten ein Enterprise-Produkt vor dem MVP — und halfen, die Retention zwei Jahre nach Launch um +20% zu steigern.



User story mapping run with dev, QA and PO; jobs framed with kill-criteria before any pixels — so scope decisions are argued, not felt.User Story Mapping mit Dev, QA und PO; Jobs mit Abbruchkriterien vor jedem Pixel — Scope-Entscheidungen werden begründet, nicht gefühlt.
Prototypes in working code (63 live screens) and a Figma design system built by AI plugins I shipped with Claude Code — hours instead of weeks.Prototypen in funktionierendem Code (63 Live-Screens) und ein Figma-Designsystem aus AI-Plugins, die ich mit Claude Code gebaut habe — Stunden statt Wochen.



Method matched to question — moderated tests, first-click, split-tests, surveys, diary-style feedback. 800+ participants, and I read my own dashboards (SQL, Python).Methode passend zur Frage — moderierte Tests, First-Click, Splittests, Umfragen, Diary-Feedback. 800+ Teilnehmende, und ich lese meine Dashboards selbst (SQL, Python).


From proposal to adoption: tokens, 50+ documented components, −40% UI implementation time — and executable systems where screens regenerate from components.Vom Vorschlag zur Etablierung: Tokens, 50+ dokumentierte Komponenten, −40% UI-Implementierungszeit — und ausführbare Systeme, in denen sich Screens aus Komponenten regenerieren.



A rare combo: product design + data science. I design with research and read the numbers myself — SQL, Python and ML literacy are part of the toolkit.Eine seltene Kombination: Produktdesign + Data Science. Ich designe mit Research und lese die Zahlen selbst — SQL, Python und ML-Verständnis gehören zum Werkzeugkasten.
| Sep 2025 — todaySep. 2025 — heutefull-timeVollzeit | Senior Product Designer · CareerSide ↗ — AI job-search platform: end-to-end product design, research, brand and growth, AI-first workflowAI-Jobsuche-Plattform: End-to-End-Produktdesign, Research, Brand und Growth, AI-First-Workflow |
| Jan – Sep 2025Jan. – Sep. 2025full-timeVollzeit | Senior UX/UI Designer · Verbly.me ↗ — AI EdTech startup, Germany: end-to-end web & mobile, scalable Figma design system, research & usability tests, 63 live screens shippedAI-EdTech-Startup, Deutschland: End-to-End Web & Mobile, skalierbares Figma-Designsystem, Research & Usability-Tests, 63 Live-Screens |
| Aug 2022 – Jan 2025Aug. 2022 – Jan. 2025full-timeVollzeit | Senior Product Designer · UX Researcher · MyOffice ↗ — native mobile suite (mail, calendar, tasks, contacts): +20% retention, −35% interface issues, 95% spec compliancenative Mobile-Suite (Mail, Kalender, Aufgaben, Kontakte): +20% Retention, −35% Interface-Probleme, 95% Spec-Compliance |
| Aug – Oct 2024 · parallelAug. – Okt. 2024 · parallel | UX/UI Designer · volunteerEhrenamt · Hope for Men, Berlin — WCAG AA accessibility, 15 usability sessionsWCAG-AA-Barrierefreiheit, 15 Usability-Sessions |
| Jun 2021 – Aug 2022Juni 2021 – Aug. 2022full-timeVollzeit | UX/UI Designer · courier navigation app — native-only: offline maps, geofencing; delivery time −18% (team result)Kurier-Navigations-App — native-only: Offline-Karten, Geofencing; Lieferzeit −18% (Teamergebnis) |
| Jan – Dec 2020Jan. – Dez. 2020part-timeTeilzeit | Data Scientist & Interface Design · part-time, 30 h/weekTeilzeit, 30 Std./Woche — neural networks + the dashboards engineers read them onneuronale Netze + die Dashboards dazu |
| 2017 – 2023 | Studies, parallel to workStudium, parallel zur Arbeit — B.Sc. Applied Informatics (2017–2021) → M.Sc. Design & Usability, ITMO (2021–2023); spring 2021 = final bachelor semester and the full-time switch into designB.Sc. Angewandte Informatik (2017–2021) → M.Sc. Design & Usability, ITMO (2021–2023); Frühjahr 2021 = letztes Bachelor-Semester und der Vollzeit-Wechsel ins Design B.Sc. Applied Informatics (2017–2021) — Data Science focus: the engineering foundation behind my SQL, Python and ML literacy; the university is in the top 300 of the QS regional ranking (Emerging Europe & Central Asia) ↗. M.Sc. Design & Usability, ITMO University (2021–2023) — HCI, research methods, prototyping. ITMO University is listed in Anabin ↗, Germany's official database of recognized foreign institutions — the degree counts in Germany — and ranks among the world's top 1000 universities in the QS World University Rankings ↗.B.Sc. Angewandte Informatik (2017–2021) — Schwerpunkt Data Science: das technische Fundament meiner SQL-, Python- und ML-Kenntnisse; die Universität gehört zu den Top 300 im regionalen QS-Ranking (Emerging Europe & Central Asia) ↗. M.Sc. Design & Usability, ITMO University (2021–2023) — HCI, Research-Methoden, Prototyping. Die ITMO University ist in Anabin ↗ gelistet, Deutschlands offizieller Datenbank anerkannter ausländischer Hochschulen — der Abschluss zählt in Deutschland — und gehört im QS World University Ranking ↗ zu den Top 1000 Universitäten weltweit. |
"During almost my three years working with Elina, I was consistently impressed by her exceptional skills as a senior UX designer and researcher. She often handled the most complex challenges of her team — and did so exceptionally well. … I strongly encourage Elina for a senior product designer position."
"Her expertise in all stages of the design thinking process, from empathizing with users to prototyping and testing solutions, was evident in the high-quality work she delivered. … Her designs were not only visually appealing but also technically feasible and aligned with business goals."
| The barDie Messlatte | My evidenceMein Beleg |
|---|---|
| Native app expertise, 5+ yrs | 30+ native features end-to-end for an enterprise mobile suite (mail · calendar · tasks · contacts, iOS & Android, 100+ spec'd screens, design system with tokens) + a native courier app + Verbly mobile. Platform patterns are my daily bread: HIG, Material, native navigation and motion. Calendar case → |
| Leads with discovery, not execution | Every project starts with falsifiable hypotheses and kill-criteria. Cheapest-test-first is how I work: before building a task module we split-tested two concepts unmoderated with 233+ users — and cut a "must-have" feature the data killed. Tasks case → · Hypothesis canvas → |
| Analytically sharp, research-led | I read my own dashboards — SQL and Python are in my toolkit, no translator needed. Method matched to question: CustDev with 12 C-levels, surveys n=241/252, first-click split-tests, diary-style feedback loops. Research case → |
| Communicates with influence | Presented research to top management and whole-company audiences; my ex-Head of UX wrote I "handled the most complex challenges of the team exceptionally well" and helped build the UX culture. References → |
| Thrives cross-functionally | Full trio partner: workshops with PM/engineering/analytics/legal, engineers in early, design QA after handoff — 95% spec compliance, 37 implementation issues triaged on one epic alone. See process → |
| Accessibility to EU standards | WCAG 2.1 AA in shipped work (color contrast, keyboard & screen-reader flows, 44px targets) and fluent in the EU regulatory layer: European Accessibility Act (mandatory for e-commerce since June 2025), EN 301 549, BITV 2.0 in Germany. Accessibility as design input, not an audit afterthought. WCAG in practice → |
| Puts AI to work | Not "curious about AI" — running on it: research synthesis, UI generation, prototypes in working code (63 live screens), AI design QA. AI-assisted shopping and product experiences are where I want to explore next. AI-first case → |
| Discover | Falsifiable hypotheses with kill-criteria · interviews & surveys · the cheapest test for the riskiest assumption firstFalsifizierbare Hypothesen mit Abbruchkriterien · Interviews & Umfragen · der günstigste Test für die riskanteste Annahme zuerst |
| Define | JTBD & user story mapping · MoSCoW scope · metrics tree with a North Star before any pixelsJTBD & User Story Mapping · MoSCoW-Scope · Metrik-Baum mit North Star vor jedem Pixel |
| Design | Flows · wireframes · design systems with tokens · hi-fi to platform guidelines · UX copy without shame patternsFlows · Wireframes · Designsysteme mit Tokens · Hi-Fi nach Plattform-Guidelines · UX-Texte ohne Shame-Patterns |
| Test | Moderated rounds + unmoderated first-click · A/B and concept split-tests · iterate on numbers, not opinionsModerierte Runden + unmoderierte First-Click-Tests · A/B- und Konzept-Splittests · Iteration auf Zahlen, nicht Meinungen |
| Ship | Spec with all states & edge cases · engineers involved early · design QA after handoff (95% spec compliance)Spezifikation mit allen States & Edge Cases · Entwickler früh einbezogen · Design-QA nach Handoff (95% Spec-Compliance) |
I meet stakeholders as a peer: I ask about the business first, push back with respect when the evidence disagrees, and take ownership of outcomes — not just screens.Ich begegne Stakeholdern auf Augenhöhe: erst das Business verstehen, respektvoll widersprechen, wenn die Evidenz dagegen spricht — und Verantwortung für Ergebnisse übernehmen, nicht nur für Screens.
Design that doesn't move a metric is decoration. Every project here starts with research — interviews, surveys, data — and ends with a number that changed.Design, das keine Metrik bewegt, ist Dekoration. Jedes Projekt hier beginnt mit Research — Interviews, Umfragen, Daten — und endet mit einer Zahl, die sich verändert hat.
Never "because it looks nice." I argue each choice in plain language — what it does for the user, what it does for the business, and what I'd test to prove it.Nie „weil es schön aussieht". Ich begründe jede Entscheidung verständlich — was sie für Nutzer bewirkt, was fürs Business, und wie ich es testen würde.
Teams should never wonder where things stand: regular status updates, transparent progress, polished handoffs at every stage, and deadlines that hold.Teams sollen nie rätseln, wo etwas steht: regelmäßige Updates, transparenter Fortschritt, saubere Übergaben in jeder Phase und Deadlines, die halten.
A product built with no static mockups at all — every screen designed directly in code with AI, 63 of them live.Ein Produkt ganz ohne statische Mockups — jeder Screen direkt im Code mit AI gestaltet, 63 davon live.
AI-first case: Verbly →AI-First-Case: Verbly →AI as a co-worker, not a replacement: it built the Figma design system through plugins while I directed every design decision.AI als Co-Worker, nicht als Ersatz: Sie baute das Figma-Designsystem über Plugins, während ich jede Designentscheidung führte.
Figma built by AI plugins: blogi →Figma aus AI-Plugins: blogi →I don't hand design over to AI — I use it deliberately, as an assistant with a clear role. I set the direction, make the decisions and own the quality; AI accelerates the work around them: research synthesis, UI variants, turning ideas into working prototypes. My ML background means I know what the models can and can't do — and when not to trust them.
Ich delegiere Design nicht an AI — ich setze sie gezielt ein, als Assistentin mit klarer Rolle. Richtung, Entscheidungen und Qualität liegen bei mir; AI beschleunigt die Arbeit drumherum: Research-Synthese, UI-Varianten, Ideen als funktionierende Prototypen. Dank ML-Hintergrund weiß ich, was Modelle können, was nicht — und wann man ihnen nicht trauen darf.
$ research # synthesize 56 interviews with LLMs — days, not weeks $ ideate # AI-generated flows & UI variants, curated by me $ build # hi-fi mockups & working prototypes in code $ test # unmoderated studies, auto-analyzed $ ship # hand-off specs devs actually follow — then measure & iterate