whoami

by Rashard · IT Security Analyst & AI Risk Practitioner

IT Security Analyst bridging the gap between AI experimentation and enterprise security. My work focuses on AI Risk Architecture, enforcing secure-by-design principles in LLMOps deployments, and building out AI Data Loss Prevention (AI-DLP) guardrails.

Focus Areas

• AI Governance & LLMOps: Managing the lifecycle, RBAC, and secure deployment of enterprise AI agents.
• AI Data Loss Prevention (AI-DLP): Configuring classification modules to track and block PII (SSN, DL, PDFs) egress in generative AI prompt flows.
• Security Automation: Designing human-in-the-loop SOAR workflows.
• DevSecOps (In Progress): Transitioning toward infrastructure-as-code via Terraform and pursuing the Anthropic Claude Architect Certification.

Repo Layout

• now.md — what I'm working on this month
• /writeups — generalized technical writeups
• /decisions — architectural decision records

Working Principles

• Smallest thing that solves the real problem
• Threat model before automation
• Documentation is a deliverable, not an afterthought
• One percent better daily

Builds

RRA Automation Pipeline — Human-in-the-Loop SOAR Pattern

A repeatable 5-step pipeline for Rapid Risk Assessment workflows built on Google Sheets + Zapier + Slack.

Core Philosophy: Automation stays dormant while a human works. A single status cell change (Automation Status = Ready to Send) is the only trigger that moves data into production.

The 5 Steps:

1. Ingest — Request ticket arrives on the intake board
2. Document — Analyst fills a strict flat-row schema on the review sheet
3. Gate — Zapier filters for Ready to Send rows only
4. Alert — Zapier sends a YAML block to Slack
5. Lock — Zapier writes Sent back to the row, preventing duplicate triggers

Teaching Model (Three Terms):

• The Sandbox — Tab 1: freeform human thinking and notes
• The Flight Deck — Tab 2: structured flat row for automation
• The Key — Automation Status: the explicit switch that moves data from sandbox to production

In Progress: Parallel variant using Zapier MCP so an LLM agent can natively trigger this workflow via Model Context Protocol instead of spreadsheet polling.

Stack

CrowdStrike NG-SIEM · Charlotte AI (AgentWorks) · Model Context Protocol (MCP) · SOAR · Google Workspace · Python · React · Tailwind CSS · Git