AI
B2B
Prototype
0 to 1
Impact
The new triage & investigation console 10×ed operator efficiency , cut time-to-triage from 10 to 2 minutes, reduced error rate by 12%, and secured Graphen’s first enterprise deal—a $2M contract.
I led the pivot, discovery and field validation of a novel human-in-the-loop AI threat diagnostic tool in 2 weeks. The prototype passed simulated tests in a code environment and achieved 90% detection accuracy, 40% improved efficiency and scored 4.5/5 trust sentiment from expert analysts.
Case Study Bites
User Understanding & Product Framing
Interviewed 10+ security analysts and domain experts to uncover hidden friction, mental overhead, and mismatched assumptions across roles. This work shaped three focused user types and clear task flows, which aligned the team on what mattered most. It clarified product vision, scoped the MVP, and prevented wasted effort—without it, we risked building features no one would buy.
Human-Centered Dashboard
Simplified the interface to reflect users’ mental models and skill levels. Machine learning signals were translated into accessible severity scores and visual flags that matched clients’ internal workflows. Thousands of raw alerts were bundled and prioritized with AI-driven ratings—boosting triage efficiency with 5x faster alert processing and 60% reduction in manual review load.
Triage Flow Redesign
Re-architected the triage interface to prioritize clarity and actionability. Key alerts, context, and controls were streamlined to reduce friction and support fast, confident resolution—resulting in a 10× efficiency gain.
AI Validation Tool
Designed a human-in-the-loop diagnostic tool that surfaces AI predictions with clear, explainable context. Advanced users can verify, override, or escalate. The prototype passed a simulated test with 90%+ accuracy and scored 4.5/5 in trust sentiment.
Modular Design System in Production Code
Designed and implemented a lightweight, scalable system directly in production (CSS/React), enabling consistent, accessible UI across the MVP. It improved quality and accelerated delivery within an aggressive timeline, saved 4 weeks+ development time.
