Real problems,
measurable outcomes
End-to-end case studies spanning AI agents, accessibility, incident management, and data visualisation — all shipping to real enterprise users.
Enhancing Visualization of Dashboards
Tokenized, AA-compliant, CVD-aware palette for security dashboards. JND analysis, CVD simulations, and A/B testing to make dense data readable for everyone.
AgenZ — Building Trustworthy AI Agents in Minutes
Security operations blocked by repetition and tribal knowledge. AgenZ lets analysts create safe, reliable AI agents without technical expertise — automating routine tasks while accelerating incident response.
Breaking Down Regex Barriers
Threat feed ingestion into Chronicle was slow, technical, and error-prone. Redesigned to visual-first for non-technical SOC analysts.
Incidents — From Queue to Closure
Analysts resolving only ~7 ActOns/day due to workflow friction. We mapped real workflows and introduced human-in-the-loop AI to improve prioritization and throughput.
AI Skins — Beyond Chatbots
Contextual AI layer — Inline Assist, Chart Insights, Smart Triggers — that adapts to the user's context instead of demanding attention.
The Pipeline That Had No Memory
Designed a video transformation module for AI training data ops — replacing FFmpeg CLI, VLC, and Slack threads with an auditable, non-destructive pipeline. Including the V1 failure that taught me the most.
