Day 1 — The Clean Handoff

Your first morning on the American Express data platform migration starts with a 9:00 AM Zoom call from your home office. You joined Amex's Enterprise Technology group last week and today is your formal project start. Priya Sharma, your manager and the VP of Data Platform Engineering, is already on camera when you join. Behind her, a home office bookshelf is lined with O'Reilly data engineering books and a framed photo from a Snowflake Summit keynote.

"Welcome to the programme. I'm going to give you the real version," she says, adjusting her headset. "This migration has been approved, funded, and staffed. The architecture is solid — Marcus Thompson designed it, and Snowflake Professional Services validated it. Deloitte is onboarded and ready. On paper, this is a straightforward lift from Teradata to Snowflake with an ETL modernization layer. Twelve months, $9.2M, four data domains, 400 Informatica jobs to convert."

She pauses and pulls up a Confluence page — the project's master documentation hub. It's clean, well-organized, clearly her team's work. "The business case is airtight. Teradata costs us $3.8M a year in licensing alone. The hardware is EOL in 16 months. Snowflake gives us elastic compute, modern tooling, and a path to real-time analytics. Ravi Kumar — our CIO — approved this because the economics are undeniable."

She leans forward slightly. "But I want to set expectations. This is financial services. Every byte of cardholder data is PCI-DSS regulated. Every financial reporting transformation is SOX-audited. Tom Fitzgerald in Compliance has a checklist that will make your eyes water. And 200 analysts depend on dashboards that better show the exact same numbers after migration as before. The technology is proven. The regulatory and organizational complexity is what makes this hard."