He mentioned there was a lot of customer facing issues to tackle, it was a startup, and their team had 4 devs. Scale shouldn’t be an issue until you know you’ll scale. Who cares if you use Kafka and Kubernetes. You sacrifice so much man power to do things thinking you’ll have a million users, but then fail like 90% of all other startups. I’ve never once heard of a startup failing because they had trouble scaling their architecture — just their business. We know he lost his job because the CEO valued the perspective of the ex googler, but the ex googler had 5 years on systems that were already large, massive man power, and budgets that dwarf startups. You’ll make any new CEO salivate showing road maps that will handle 1,000,000 concurrent users, but “handle” and road maps to “get to” are different.