SHARKPARTY
Case studies
Case study · Frontend · ML · Design Systems

IdentityAI.

SailPoint IdentityAI embeds machine learning into identity security—surfacing risky and anomalous access so enterprises can hold least-privilege at scale. I built the interfaces and data visualizations that turned that model output into something a security team could read, trust, and act on.

Role—Staff UI Engineer, SailPoint Product—sailpoint.com
AngularD3.jsJavaScriptAWSMachine LearningData VisualizationAccessibilityDesign Systems
// The product

Least privilege, when there are too many identities to reason about by hand.

Cloud migration, work-from-anywhere, and app sprawl exploded the number of identities and access points enterprises have to govern—well past what human-centric review can keep up with. IdentityAI applies machine learning (and, later, generative AI) to that firehose: access insights from attributes, roles, history and entitlements; access recommendations from peer-group analysis; and role modeling that keeps access right as the org changes. The hard product problem underneath it is trust—a recommendation nobody understands is a recommendation nobody follows.

// My part—the layer humans actually touch
01
Made ML legible

Built the interfaces that let customers interpret machine-learning-driven insights across identity data, entitlements, access history, and user activity—turning model output into decisions a human could actually make.

02
Risk, visualized

Developed dashboards and interactive visualizations for anomalous access behavior, risk scoring, and contextual identity-governance insights.

03
D3 in focus

Created dynamic D3-based charts to represent suspicious access patterns and near-real-time anomaly-detection signals—the kind of viz a security team lives in.

04
Accessible by default

Designed and implemented accessible UI components that integrated with the backend identity-analytics systems.

05
A shared system

Contributed to a reusable design system that improved UI consistency, accessibility, and development velocity across product surfaces.

// This is where identity and I first got properly entangled—the start of nearly six years in the space.

// What the product delivers
40%
Faster to detect & respond
2 min
To deprovision (from 30+ days)
~80%
Breach-cost savings, full AI + automation
11k hrs
Saved by a SailPoint customer

// Figures published by SailPoint for IdentityAI (2024). My contribution was the frontend and visualization layer during its early days (2017–18)—not these later customer outcomes.

// Want the deep version?

Happy to talk through the hard parts.

Email Kam More work