Infragistics' core customers were developers relying on UI components for built-in functionality and performance. A pattern kept surfacing on customer calls: after all the design work, engineers were still rebuilding pages by hand from static specs. Designers wanted to deliver finished UIs using real components, not handoff documents. Any solution had to span Angular, React, Blazor, and Web Components.

I owned product strategy, roadmap, and design execution, influencing without authority across 20 people in 4 distributed teams in Uruguay and Bulgaria: developers, designers, and architects covering core UI, design system, Figma integration, and cross-platform code generation. Reporting to the director of engineering; sprint sizing with a development product owner. My role sprint by sprint: spec, test, sign off, and own the customer relationship: sales calls, requirements capture, roadmap decisions.

Nine months from vision to v1, with a newly assembled remote team and no shared design language yet in place. The technical constraint compounded it: enterprise stacks shift, and tying designs to any one framework would create rework the moment they did. Scope had to be cut ruthlessly to stay coherent and hit the date: the design itself had to be the stable artifact, with code generated on demand for whatever framework a team needed.

Every call was a constraint decision. The component API was extensive. I audited it down to the properties that mattered in a visual editor, so it felt like a design tool, not a developer console. Theming was scoped to CSS overrides: maintainable, readable, no platform expertise required. The output was framework-native code, not a proprietary format; developers could extend it freely with the full API. Design system fidelity was non-negotiable: GitHub and Azure DevOps integrations put generated code directly into existing pipelines, with designers authoring PRs themselves.

When the AI layer was added, the constraint architecture became the trust architecture. A component metadata layer gave the LLM a constrained vocabulary: hallucinations became mapping errors, not invented components. Visual scope became the change boundary. Testing surfaced two mental models: blank-slate generation and iteration on existing work. Each maps to a separate instruction file. Fully autonomous agents remain the open problem.