[ ] The "Toy" Filter: My top 3 projects do NOT use Titanic, Iris, or MNIST datasets (Recruiters now auto-reject these as "homework").
[ ] The Messy Data Test: At least one project uses data I scraped or found myself with $>20\%$ missing or corrupt values (showing I can handle "real-world noise").
[ ] The Production Check: I have one project that is "live" (using a tool like Streamlit or a cloud API) rather than just a static Jupyter Notebook.
[ ] The CEO Pitch: I can explain my best project in 60 seconds without mentioning a single library (e.g., "This model reduced marketing waste by £15k").
[ ] The ROI Metric: Every project on my resume includes a quantified financial or operational result, not just "99% accuracy.".
[ ] The $P > 0.05$ Test: I can explain a confidence interval or statistical significance to a non-technical manager in plain English.
[ ] The AI-Native Edge: I use professional AI editors (like Cursor IDE) to architect entire systems, not just to generate snippets of code.
[ ] The Agentic View: I have a basic understanding of "AI Agents" and how they are replacing routine data cleaning tasks.
[ ] Prompt Mastery: I use the Role-Context-Goal framework to get high-fidelity outputs from LLMs for business analysis.