Part 7 of 10 · 2 min read
Mistakes that sink a design interview
Common mistakes, and what the example does not do.
Mistakes that sink a design interview
- Drawing boxes without requirements or numbers.
- Designing for servers when tokens are the bottleneck.
- Synchronous requests for minutes-long runs.
- Ignoring tenant isolation, or leaving it to "the model will not do that".
- No durability story: a worker kill restarts everything.
- Full autonomy on day one for money-moving actions.
- No cost cap, no eval, no rollback plan.
- No failure modes listed, or only provider outages.
- Over-engineering: Kubernetes, many agents and a vector database before a single working loop.
- Not saying what you would build first.
What the example does not do
About this example What the example does not do
- No system was built or load tested; the numbers are estimates from assumed inputs.
- Tokens per run assume full history re-send and ignore compaction, caching hit rates below 100 percent of prefixes, and provider-specific accounting (many providers count cached tokens differently against rate limits).
- Costs use illustrative prices, not a current price list, and ignore infrastructure, storage, engineering and human review costs.
- Little's law gives averages; real queues have variance, bursts and retries.
- The rubric is the author's own, not an official standard; real reviews depend on context.
- The two designs are sketches, not reference architectures; they were not validated with a real team.
- Optional real experiment: take your own agent, measure real steps, tokens and run times from traces, replace the assumptions in
design.pyand compare the predicted cost and concurrency with a day of real traffic.
Check yourself
You are asked to design a support-agent platform. What do you do first?
Show answer
Clarify requirements, then estimate. For 200 tenants at 500 tickets a day that is 100,000 runs a day, about 1.2 a second on average and 5.8 at peak, and with 40 s runs about 230 running at once at peak.
For the curious A design answer names its failure modes
Name a control for each: provider outage (breaker and fallback), worker kill (journal), noisy tenant (quota), runaway loop (cap), bad release (canary and rollback), prompt injection (Level 10), data leak (isolation and erasure), and cost spike (alerts and caps). All workload and price numbers are assumed round figures; the practice is the structure of the conversation.
What you will use daily at work
| Practice | What it looks like |
|---|---|
| Estimate first | Little law: concurrent runs equal arrival rate times run time. |
| Walk one request end to end | Then go deep on the two or three riskiest parts. |
| State trade-offs and what you would not build | Say why. |
| Check your design against the full list | The rubric in the example code lists the 14 areas. |