Mistakes to Avoid

BAD: “I’d just add more shards.” – Shows focus on scaling without addressing latency SLAs. GOOD: “I’d add shards, then instrument read latency, and create a run‑book for fallback if 95 % of reads exceed 5 ms.”

BAD: “Enable all CloudWatch metrics.” – Cost‑blind, triggers “Dive Deep” failure. GOOD: “Enable key metrics, set alerts for > 100 ms, stay under $3 K budget by sampling at 1 %.”

BAD: “I’d A/B test the fix.” – Lacks operational hand‑off, leads to “No Hire” in Google debrief. GOOD: “Deploy canary, monitor SLO breach, roll back via automated run‑book.”

---

FAQ

Why does Google penalize a candidate who mentions only shard count? Because the debrief panel (6 interviewers, 5‑1 Y) flagged the answer as missing latency SLA consideration; reliability outweighs raw capacity.

Can I succeed at Amazon without a cost‑aware answer? No. In the Q2 2024 loop the candidate who said “spin up a larger EC2” lost after a 2‑2 split; senior manager Karen Wu cited cost blindness as the deal‑breaker.

What compensation should I expect if I get an offer? Google SRE L5: $185 000 base, 0.07 % equity, $30 000 sign‑on. Amazon SRE L6: $165 000 base, $40 000 RSU, $20 000 sign‑on.

---amazon.com/dp/B0GWWJQ2S3).