Start with the constraint
I make the deadline, technical boundary, accessibility requirement, and user risk visible before choosing a direction.
Mateo Cohen
Lead Product Designer and AI Strategist
I turn ambiguous enterprise briefs into shipped products. My recent work spans generative AI platforms, fintech onboarding, and university-scale systems, usually as the sole or lead designer working directly with engineering. I design in code as well as Figma, run my own research, and teach teams how to work with AI safely at production speed.
How I make decisions
Use convention deliberately
I reuse proven patterns by default, then deviate only when research shows the standard is failing the user.
Separate AI from certainty
I assign deterministic work to code, probabilistic work to models, and keep people in control where errors carry consequences.
Core skills
- AI product design
- AI strategy
- Agentic workflows
- Human-in-the-loop AI
- Enterprise UX
- Design systems and tokens
- Accessibility
- UX research
- Design engineering
- Rapid prototyping
- Cross-functional leadership
Selected work, with outcomes
Commercial card onboarding redesign
Lead UX Designer on a digital onboarding experience that was bleeding time. I ran the moderated research myself, made accessibility part of the definition of done, and spent innovation only where testing proved standard patterns failed.
- Handling time cut 65%, from 54 days to 20
- 96% of the pipeline migrated to digital
- 98% of users rated the flow easy or very easy
- 20,000 annual hours of productivity unlocked
Faculty Information System redesign and AI pipeline
Lead Product Designer on a university-wide platform with seven years of usability debt. A three-month contract expanded to six as the work grew to include redesigned navigation, one consolidated design system, and an AI pipeline that ended manual CV data entry with a human review step on every extracted field.
- 85% first-pass parsing accuracy on 20 to 40 page CVs
- 75% reduction in admin review time
- Redesign shipped to production; AI track earned an Amazon pilot for 2026
Generative AI contract intelligence platform
Sole product designer on a zero-to-one GenAI platform that processes 30,000 to 40,000 pages of provider contracts per engagement. I built the product plan from an ambiguous brief, designed in Figma and code on a four-person delivery team, and left an agent-powered roadmap the team is still building.
- Concept to production in 5 months, shipped on deadline
- 600+ production iteration commits across the build
- Engineering contract renewed as the team continued the roadmap
Security platform redesign
Led a cybersecurity platform redesign informed by 40 interviews, threat-visualization research, and a unified design system.
- Six months from research to shipped v6 design
- Component library adopted for the v6 platform
- Redesign used as a lead artifact in client demos
Conversational AI platform evaluation
Evaluated more than 50 conversational AI platforms, built working prototypes, and created an evaluation framework that supported a Phase 2 investment decision.
- 50+ platforms evaluated and narrowed to 5 finalists
- 2 comparable working prototypes in Watson and Dialogflow
- $1M+ Phase 2 investment approved after prototype review
Full case studies, videos, the interactive journey, resume, and contact details are passcode protected to respect confidential client work. Authorized reviewers and recruiters can request access through the portfolio entry page.
Enter the portfolio