Josué Zeledón

Large-scale betting platform migration — from external agency to internal UX practice. As UX Senior, I led the end-to-end UX process: research, content strategy, information architecture, wireframing, prototyping, usability testing, and a complete design system built from scratch with Rocket Air. The result: a cohesive modernized experience, measurable growth in visits, and UX becoming an internally trusted driver of product decisions.

When this project began, the internal UX team was newly formed and still earning trust — much of the product strategy had been defined by an external agency. The migration spanned multiple product areas with evolving business requirements and high technical constraints. Collaboration was continuous and cross-functional, involving workshops, design reviews, stakeholder sessions, and feedback loops — not handoffs. This approach gradually shifted confidence from the agency to the internal team.

The company inherited a complex betting platform with inconsistent interfaces, unclear ownership, and limited documentation. The lack of shared standards and alignment between design, product, and engineering made it impossible to maintain consistency, meet evolving business needs, or deliver efficiently across product areas. UX had no seat at the table — and no credibility yet.

From a legacy platform with inconsistent interfaces to a modernized, cohesive experience with internal UX ownership.

Growth in visits confirmed by post-migration SEO data. Platform migrated with cohesive and modernized UX. Internal team gained full UX ownership — replacing the external agency. BA Core design system now powers consistent design across multiple verticals and product lines.

Trust is built incrementally — and UX credibility is earned through consistent delivery, clear communication, and documented decisions, not just great design. The shift from agency-led to internal ownership required as much organizational work as it did design work.

What I'd do differently: establish documentation habits from day one. Every workshop, every decision, every iteration — documented. Well-documented artifacts are what continue to create value long after the project ends.

I contributed to the maintenance and enhancements of Cisco's Design System, ensuring coherence and accessibility across multiple platforms. I created presentation templates prioritizing storytelling — weaving research narratives and findings to communicate design decisions to global stakeholders.

Contribute to an enterprise design system used across thousands of pages and dozens of product teams worldwide — where each enhancement had to be backward-compatible, accessibility-compliant (WCAG AA), and approved by global stakeholders before touching a single component in production.

Component enhancements successfully published to Cisco's global design system library, now active across multiple product lines. Presentation templates changed how the team communicated research to stakeholders — reducing ambiguity in decisions and accelerating approval cycles across time zones.

At Accenture, I built a Figma prototype to demonstrate the usability of next-generation SynOps software. I designed a highly customizable luxury car buying simulation — including VR test drive, vehicle configuration, and metaverse commerce.

Build a compelling demo prototype under 2 months — showing how next-generation AI software could transform a complex, high-stakes purchase (luxury vehicles) into an immersive, personalized experience. The prototype had to feel production-ready for C-suite stakeholder presentations at Accenture.

Demo prototype approved and presented in live Accenture SynOps stakeholder sessions. The experience became a reference for how next-generation commerce could look — combining VR, AI personalization, and metaverse integration into a single cohesive UX narrative that stakeholders could experience, not just observe.

Design a music creation app for a multi-touch table — a non-standard interface with unique constraints: large surface, no cursor precision, potential simultaneous multi-user interaction, and no physical controls. The audience: musicians of all levels who shouldn't have to think about the interface while creating.

App for amateur and professional musicians. Create, compose, record, and share music through a modern interface. I designed the complete custom iconography and app logo.

Semantic phrase: MODERN & INSPIRING — a modern interface, as simple as possible to promote creativity.

PULSE demonstrated that a complex multi-touch music interface could be both powerful and approachable. The semantic guiding phrase proved its value — every ambiguous decision had a test. A formative project that shaped my understanding of interaction design for non-standard surfaces and constraint-driven creativity.

EliteMarket is a premium digital services marketplace built for the Costa Rican market — connecting verified local sellers (Marketing, UX/UI, Audiovisual) with buyers through a trust-first, mobile-first experience. I designed the entire product from zero: architecture, design system, component library, and user flows — with Claude and Figma MCP as integrated AI partners throughout.

Costa Rica's digital services market has a trust problem. Buyers don't know if sellers are real, and sellers lack a premium channel to showcase their work. The core UX challenge: make quality, safety, and discovery feel effortless — on mobile, for a market that lives on WhatsApp.

This project was an experiment in working with AI not as a tool, but as a collaborator. Every major design decision had Claude as a sounding board. The workflow changed how I think about velocity, documentation, and the boundary between design and implementation.

The token system follows three layers — from raw values to component-level decisions — synced live between Figma and code. No hardcoded values anywhere in the codebase.

Working with AI as a design collaborator — not just a code generator — fundamentally changed my process. I made better-reasoned decisions faster. Documentation happened in real time. The distance between 'design intent' and 'shipped code' collapsed. The skill that evolved most wasn't using AI — it was knowing which questions to ask it.