Gunther Eisen (Gonçalo Ferraz)
Senior UX Leader · Principal UX Consultant
Curitiba, Brazil — Remote-Ready for Global Opportunities
I help product teams understand what users actually need, and why the current experience falls short of that.
About
Senior UX Leader & Principal Consultant
20+ years of UX practice at the intersection of research, strategy, and design. My work spans qualitative and quantitative research, information architecture, usability evaluation, and product strategy — at principal level, for teams that need more than a report.
I have led engagements for B2B SaaS platforms, legal systems, web hosting products, health tech, and financial services — working embedded with product teams and as an independent consultant for digital agencies. The common thread: making the gap between what users need and what exists impossible to ignore.
I co-founded Faber-Ludens Institute, Brazil's first Interaction Design school, where I directed operations and taught for over a decade. I have trained and mentored hundreds of UX and product professionals — and that perspective on building capability, not just delivering findings, shapes how I work with every team.
Experience
Senior leadership across SaaS, fintech, and health tech
— Present
— Present
— 2025
— 2021
— 2020
— 2018
Education & Credentials
Academic background and professional recognition
-
MA, Education (Coursework Completed)PUCPR · 2015–2016 · CAPES Scholar
-
Certificate, Creative Direction with Digital MediaUC San Diego · 2000–2002
-
Graduate Diploma, MarketingPUCPR · 1999
-
Bachelor of Industrial Design, Graphic DesignUNOPAR · 1995–1999
-
Analytics & UX · Lean UX & Agile · Design LeadershipNN/g Nielsen Norman Group · 2020–2021
-
Rapid UX Research for Lean & Agile TeamsNN/g Nielsen Norman Group · 2020
-
C1 Professional English ProficiencyETS TOEIC · 955/990 · 2013
-
Co-author: Web Accessibility GuideW3C Brasil
-
15+ keynote & panel talksCampus Party · World Usability Day · FrontInterior
AI Practice
AI-native UX practice, built and running in production
I don't use AI tools — I architect AI systems. Over the past year I built a fully operational AI-native UX practice: a governed multi-agent workforce, automated research-to-delivery pipelines, and local ML infrastructure running in daily production. The result is a practice that closes every manual bottleneck in the UX workflow — from research ingestion to client handoff.
Multi-agent orchestration
20+ specialized agents operating as a governed workforce — each with a defined role, formation history, and process assignments. Parallel specialist reviews, cross-agent delegation chains, and equipment checks before new project types. No human mediating technical detail between agents.
Research-to-delivery automation
Full pipeline from raw inputs to shipped deliverables: 131-hour scope estimate generated from a PRD + meeting transcripts + email threads; 44 GitHub Issues created without manual copy-paste; 6 IA deliverables produced with zero human authoring. Multi-locale UX audits across 6 simultaneous locales.
Production ML infrastructure
Local vector search engine (ONNX, Apple MPS) powering dual-corpus knowledge bases — classical library and professional domain KB. Multi-fallback PDF ingestion pipeline. Memory hygiene automation via local LLM at zero API cost. Not a demo — runs in daily production.
Custom MCP server development
Built a purpose-designed RSS MCP server (Python + FastMCP + SQLite) for agent intelligence sweeps. Integrated and configured 12 MCP servers across design, deployment, research, and communication layers.
Spec-to-prototype pipeline
Wireframes generated from IA screen trees via Stitch MCP. HTML/CSS prototypes deployed to Cloudflare Pages in a single command. NotebookLM knowledge bases delivered to clients as self-service Q&A — the AI layer becomes a client deliverable, not just an internal tool.
AI governance design
Token economics as a first-class operational constraint. Modular instruction loading — 75% reduction in always-loaded context. Agent council model for knowledge base curation. Strategic review gates as real approval checkpoints. Organizational design principles applied to AI systems.
Stack: Claude Code CLI · Perplexity MCP · Firecrawl MCP · Figma MCP · Stitch MCP · Excalidraw MCP · Cloudflare Pages + Wrangler · Gemini 2.5 · NotebookLM · Ollama · SentenceTransformer ONNX · FastMCP · GitHub CLI · Notion MCP
Contact