Work at Kontent.ai

End-to-end design of a conversational AI agent for enterprise content management – enabling content teams to automate complex tasks through natural language.

• Complete chat interface: welcome states, input patterns, message rendering
• Thinking indicators showing AI reasoning progress through multi-step tasks

Embedding-based semantic search integrated into enterprise CMS – enabling content editors to find relevant content using natural language.

• Semantic search integration into existing Content Inventory UI
• "Whisper/Suggest" component – prompt-engineered suggestions teaching users effective semantic query patterns

AI-powered text generation tools integrated into the rich text editor – enabling content creators to generate, refine, and transform content using natural language.

• Make shorter / Make longer – Adjust content length while preserving meaning
• Change tone – Adapt writing style for different audiences

Audit and tracking system design for AI actions in enterprise CMS – implementing the Accountability principle from the Responsible AI Framework to provide clear visibility into AI vs. human actions.

• External AI integrations (third-party LLMs via API)
• Internal AI features (Ask AI, WYSIWYG AI)

AI-powered content translation enabling content teams to translate items across languages using large language models.

• Led retention research analyzing post-launch adoption patterns
• Conducted user interviews to understand translation workflows

AI-powered asset suggestions and duplicate detection system for enterprise asset management – helping teams avoid duplicate uploads and surface relevant existing assets.

• Background duplicate detection during upload (exact binary + semantic similarity)
• Asset library analysis to identify existing duplicates

Inline AI instruction system integrated directly into the rich text editor – enabling users to write prompts that expand into generated content without leaving the editing context.

• Incomplete (unconfirmed) prompt – User is still typing the instruction, system waits for confirmation before processing
• AI result awaiting approval – Generated content is displayed but needs explicit user confirmation before becoming standard text

Visual highlighting system for AI-generated content – enabling content teams to easily identify and review AI contributions within their content.

First designer at Kontent.ai to work directly in the codebase – pioneering "designers own the frontend" approach using AI-assisted coding tools.

• Modified Card component to support optional icons with configurable spacing
• Applied inheritance patterns affecting 53+ component instances

Multi-factor authentication management for enterprise users – the first project where a designer built the entire frontend, establishing the "designer owns frontend" model at Kontent.ai.

• Designed complete MFA setup and management flows
• Created state management UI (enabled/disabled/enforced states)

Strategic initiative establishing ethical AI principles and governance framework across the product organization – providing shared language for evaluating AI features.

• Fairness – Ensuring AI features work equitably across diverse content types and user scenarios
• Human Agency & Control – Keeping humans in decision-making loop with clear approval flows

Systematic framework for measuring feature success after launch – combining quantitative analytics with qualitative research.

• Define metrics before launch – What does success look like?
• Set up tracking – Consistent event taxonomy in Amplitude

Design system guidelines establishing baseline consistency across all AI features – ensuring users have a predictable and trustworthy experience with AI tools throughout the product.

• Use Simple Status – Consistent communication of AI system status with spinner during interaction, alert type for errors, success type when finished
• Use Consistent Actions – Standardized actions (Cancel, Try again, Change instruction during interaction; Dismiss, Try again, Change instruction for errors; Approve, Discard, Try again, Change instruction when finished)

Managing and optimizing the design process at Kontent.ai through codification, analysis, and experimentation – reducing iteration time and establishing repeatable frameworks.

• Efficiency and effectiveness patterns across different project types
• Bottlenecks in the design-to-development handoff

Ongoing maintenance and governance of Kontent.ai's design system across Figma and Storybook – ensuring consistent UI and reducing design-to-development friction.

Collaborative problem-solving framework encouraging cross-functional teams to explore solutions before detailed design begins – significantly improving project kick-off efficiency.

• JTBD drill-down – Understanding jobs to be done
• User story mapping – Visualizing user journeys

AI-powered Figma plugin that automatically compares design system components with documentation – reducing manual comparison time from hours to minutes.

• Design systems have discrepancies between documentation (Kontent.ai) and implementation (Figma)
• Manual comparison takes 2-3 hours per component

Hackathon project exploring systems for rolling out beta features to select customers – validating approaches for feature flagging, customer communication, and feedback collection.

Internal hackathon project exploring AI-powered image generation from text descriptions – investigating potential applications for content teams working with visual assets.

• Validated technical feasibility of AI image generation integration
• Identified potential use cases for content production workflows

Multi-site content management system enabling enterprise customers to manage content for multiple channels within a single project.

• Help navigate across content of multiple sites
• Manage engineering operations for multiple sites

Dynamic content type system allowing elements to show/hide or become required based on conditions – simplifying complex content models without unnecessary duplication.

Multi-device preview system enabling content editors to preview content across different screen resolutions and orientations – eliminating dependency on external tools.

• For Connie (Copywriter): Seamlessly preview content in different breakpoints without third-party tools
• For Connie: Easily share preview links to collaborate with colleagues

Asset management permissions system enabling content teams to organize and control access to media assets through collections – improving security for confidential materials while maintaining easy access to shared resources.

• Kickoff & Problem definition – Aligned with PM on goals and constraints
• Shaping workshops – Cross-functional sessions to explore solution space

Feature enabling content strategists to set default values for content type elements – reducing repetitive data entry for content creators while maintaining flexibility to override when needed.

Comprehensive redesign of the content inventory screen to improve usability, flexibility, and content discovery – transforming how content teams navigate and manage large content libraries.

• 12/2024: Internal research, secondary analysis, synthesis, first concepts
• 01-03/2025: Design iterations and feedback sessions