Dara Afraz portrait
Dara Afraz
Product Strategy · Agentic AI · Design Systems

Knowledge Base Ops (KBOps): Hybrid AI Knowledge-Base with Classifier Reasoner (Mars Rover Heritage)

UX Design, HCI Research, and Prototyping

I led the user experience blueprint for a hybrid AI platform built for small-to-medium enterprises in heavy industry—sectors that historically lack intuitive, usable tools. The platform blends symbolic reasoning with statistical inference to help operators make better decisions faster. My task: make it usable, valuable, and buildable.

Challenge

Design for highly technical users in high-stakes environments, while aligning with a product still shaping its form. Initial ambiguity was high—no clear customer archetypes, workflows, or interaction models. Many stakeholders, few constraints, and a new codebase.

My Role

I embedded within the Platform Services team as the UX and HCI lead, operating across squads. I worked with PMs to shape feature definitions, with engineers to design for feasibility, and with leadership to ensure alignment. My UX process emphasized spikes, rapid prototyping, and continuous storytelling to manage ambiguity and pace MVP development.

Key Contributions

  • UX Spikes (25+) to reduce uncertainty and align teams around approach
  • 50+ interaction and feature prototypes to test, validate, and communicate
  • Wireframe system for canvas-based, hierarchical interactions (tree logic, drag-and-drop, embedded editors)
  • Knowledge curation framework to structure symbolic-AI content for downstream use
  • HCI research artifacts to guide design decisions rooted in real user needs

Design Challenges I Tackled

  • Designing mixed-granularity editing within a canvas-based UI
  • Communicating AI reasoning without overwhelming the operator
  • Building affordances for partially-formed knowledge and incomplete data
  • Orchestrating user flows that combine structured configuration with natural input

Outcomes

The work accelerated MVP development, aligned a cross-functional team around a shared vision, and enabled the company to land early customers and demos. This platform now serves as the foundation for multiple vertical solutions—spanning operations, anomaly detection, and AI-assisted workflows.