ENGINEERING / CASE STUDIES

EXPLAIN THE
DECISION.

Technical case studies about evaluating controls and engineering tradeoffs. These are illustrative, not claims about specific implementations or results at my employer.

01 / DESIGN NOTES

From signal to system.

CASE STUDY 001 · DATA PROTECTION

Reducing false positives without losing SSN coverage

Problem. Generic nine-digit pattern matching flags reference numbers along with potential customer identifiers.

Approach. Combine format checks with surrounding field context, destination risk, and verified business use; evaluate precision and misses on a labeled synthetic corpus.

Tradeoff. Contextual detection improves specificity but can miss data in unusual exports. Introduce staged audit and block policy changes with exception governance.

Experiment with policy tuning ↗
CASE STUDY 002 · DETECTION ENGINEERING

From DLP alert to evidence-based exfiltration assessment

Problem. A high-severity content match does not demonstrate that a sensitive file left the environment.

Approach. Correlate DLP control outcomes with web/email delivery, endpoint operations, identity activity, file metadata, and approved destinations.

Tradeoff. Telemetry gaps, asynchronous logging, and exceptions can change confidence. Report confirmed, attempted, and unknown outcomes separately.

Investigate simulated telemetry ↗
CASE STUDY 003 · ARCHITECTURE

Designing an exception review lifecycle

Problem. Permanent exemptions weaken controls while requiring engineers to judge business necessity beyond their context.

Approach. Identify a business owner, record scope and expiry, route approvals, review continued need, and revoke stale exceptions.

Tradeoff. Efficient approval needs auditability and least privilege without bypassing appropriate data-owner judgment.

Explore reference controls ↗