AI OS.
Full product design ownership for an AI Operating System — from early concept to detailed production UI. Covers agent management dashboards, visual workflow builders, RAG interfaces, and knowledge graph explorers.
Abstracting the abstract.
We had to create mental models for concepts that users had never seen before—like chaining LLMs or managing vector embeddings—while keeping the interface familiar and intuitive.
Agent Complexity
Users needed to configure AI agents with tools, memory, and prompts without writing code. The UI required progressive disclosure of advanced parameters.
Workflow Visualisation
Chaining agents and tools into pipelines needed a node-based interface that was spatial, intuitive, and capable of handling infinite canvas scaling.
RAG Studio UX
Uploading, chunking, and querying documents had to feel simple despite deeply technical processes happening in the background.
Information Density
Dashboards needed to surface real-time agent status, costs, latency, and performance without cluttering the screen.
Core Modules
Agent Hub
Central workspace to create, configure, and monitor AI agents with tools, memory, and prompt injection.
Workflow Builder
Visual node-based editor for chaining agents, APIs, and tools into complex multi-step AI pipelines.
RAG Studio
Upload documents, configure chunking strategies, and query your knowledge base through a clean conversational interface.
Knowledge Graph
Interactive graph explorer visualising entity relationships extracted from ingested documents and data sources.
System Architecture in Action
Agent Hub Dashboard
Central control panel for managing all AI agents.
Workflow Builder
Node-based visual pipeline editor.
RAG Studio & Knowledge Graph
Document ingestion and entity relationship explorer.
Bringing order to chaos.
Defined the entire design system and architecture from scratch.
Core modules successfully designed and handed off.
Pioneered novel UI patterns for complex AI chaining.