Start with the problem
Name the business problem before the tools. A fresher project and a platform migration both need the same opening: who was stuck, and what had to change.
If you begin with the product list, the interviewer has to guess why the work mattered.
Show one decision
Pick a single design choice and say what you gave up. Storage, latency, cost, or a simpler pipeline are all valid trade-offs if you can explain them.
Then say how you would know the choice was wrong. That is more useful than a feature tour.
Close with what you would do next
End with the next improvement, not a summary of every task. It shows you can see past the ticket you were given.
Keep a two-minute version and a longer one. Both should use the same story.