Design a privacy-safe personalization system for Stock at global scale

Problem Statement Description

Product context: Adobe is a creative, document, and marketing software company; its products include Creative Cloud, Photoshop, Illustrator, Acrobat, Adobe Express, Firefly, and Experience Cloud.

Adobe Stock serves creators, marketers, designers, agencies, and enterprise teams who need to discover licensed photos, videos, illustrations, templates, 3D assets, and increasingly AI-generated or AI-assisted creative content. The challenge is to design a personalization system that improves asset discovery across global markets while preserving user privacy, meeting enterprise trust expectations, and fitting into Adobe’s broader creative workflows.

In this interview, assume users arrive from multiple surfaces: direct Stock search and browsing, Creative Cloud apps, document and marketing workflows, team libraries, campaign planning tools, and possibly generative AI creation flows. Personalization may involve search ranking, recommendations, related assets, recently relevant themes, organization-level preferences, licensing context, and creative intent signals. The system must work across regions, languages, content types, subscription models, and enterprise governance requirements.

You are not being asked to design only a recommendation UI. You should define the product and technical requirements for a privacy-safe personalization platform: what data it can use, how consent and controls work, how personalization is computed and served, how it integrates with Adobe Stock and creative tools, and how the team would evaluate reliability, relevance, fairness, compliance, and business impact.

The experience should consider:

- Key user types, including individual creators, enterprise creative teams, marketers, admins, and contributors whose content may be recommended.

- Personalization surfaces such as search results, home feed, similar assets, project-aware suggestions, Creative Cloud integrations, and enterprise team experiences.

- Data inputs and constraints, including user behavior, search intent, licensed assets, project context, team-level signals, regional rules, consent state, and sensitive data minimization.

- Privacy, security, and compliance requirements across global markets, including transparency, opt-out controls, data retention, access control, and enterprise policy enforcement.

- APIs, data pipelines, model serving, latency, cold-start handling, multilingual support, content metadata quality, and integration with existing Stock systems.

- Guardrails for creator trust and AI ethics, such as avoiding inappropriate inference, filter bubbles, bias in asset exposure, unsafe content amplification, or misuse of customer creative data.

- Observability and evaluation, including relevance metrics, conversion/licensing outcomes, diversity of recommendations, privacy incidents, model drift, regional performance, and system reliability.

- Rollout approach, experimentation strategy, fallback behavior, incident response, and how to safely launch across consumer, team, and enterprise segments.

Your goal is to frame a technically credible product design that balances personalization value with Adobe’s obligations around professional trust, privacy, creative ownership, and global enterprise readiness.