Choose a project. Prove the work.
Start with identity, AI security, or cloud. Then take on two deeper AI engineering builds. Every project ends with a concrete portfolio package.
Pick the next proof point.
Each path ends with a visible artifact, test evidence, and a short decision narrative—not an automatic badge or a promise of employment.
Windows AD lab
Design a lab-only directory, automate test identities, and prove policy boundaries.
Prompt-injection assessment
Test a controlled AI workflow and document what untrusted text can—and cannot—change.
LLM output defense
Keep model output away from privileged SQL and verify the boundary with synthetic data.
Cloud portfolio site
Ship a static site with HTTPS and a private origin, then defend the architecture.
AI control plane
Run Ollama behind an authenticated gateway with model policy, observability, and evals.
Personal AI assistant
Build a useful Jarvis-inspired assistant that asks before acting and protects private data.
Show the decision, not just the screenshot.
Define the outcome
Write the goal, constraints, and what success looks like before you build.
Keep the receipts
Capture redacted configuration, tests, failures, and fixes in a clear README.
Tell the story
Walk through your architecture and tradeoffs in a mock interview.
Try the work.
Use the guided lab or create your own isolated environment. Never test someone else’s system without authorization.
Practice the hiring conversation.
Use the mock interview to explain decisions, validation, and the next improvement. A free account is required for the demo.
Turn the build into a story.
Present what you built, what failed, how you verified it, and what you would improve.