Scenarios

Questions that start with “What would you do if…”. There’s rarely one right answer, but there are weak, decent and strong ones.

Each scenario shows all three, and what the interviewer listens for: ownership, staying calm, using evidence, and preventing a repeat.

Looking for behavioral questions like “tell me about a conflict”? Those live in the Interview Playbook, with the STAR method and a story bank.

How to use this roadmap

• Topics are in teaching order. “Builds on” links show what to know first, even from other tracks.
• New? Start at the top. Know some already? Jump to a level, or skip any topic you can already explain.
• Short on time? Tick “Asked often only” to see what interviews ask most.
• Your progress is saved in this browser.

Best after: Python, SQL, System Design. Do the core levels of Python, SQL and System Design first. Scenarios mix them all.

16 topics in 4 levels, about 6 h of reading. With practice, plan 2 to 3 times that. 1 of 16 topics have lessons so far.

4 practice problems from this track’s lessons, in one list.

Foundations 5Core 5Advanced 5Interview-ready 1

Foundations

1. 1. How scenario answers are gradedComing soon

   What separates a weak, a decent and a strong answer.

   10 min read

2. 2. Latency spiked after a deployAsked oftenNot started

   The site got slow right after a release. What do you do first?

   • p95 latency tripled after a deploy
   Start →

   20 min read, about 1 h 20 min with practiceBuilds on: How scenario answers are graded

3. 3. The error rate jumpedAsked oftenComing soon

   Errors suddenly spike in production: calm, step-by-step triage.

   20 min readBuilds on: How scenario answers are graded

4. 4. A Python script is slowComing soon

   Find out why a script is slow before you try to fix it.

   15 min readBuilds on: How scenario answers are graded

5. 5. Memory explodes on a large CSVComing soon

   Process a file bigger than memory without crashing.

   15 min readBuilds on: How scenario answers are graded · Files and pathlib (Python)

Core

1. 1. Production: memory leak, DB CPU at 100%, deadlocks, disk full, intermittent 502sAsked oftenComing soon

   Common production fires, and how to find the cause with evidence.

   35 min readBuilds on: The error rate jumped

2. 2. Cache stampede and queue backlogComing soon

   When the cache empties all at once, or a queue keeps growing.

   20 min readBuilds on: Caching: cache-aside, write-through, write-back, eviction, TTL (System Design) · Queues and pub/sub (System Design)

3. 3. Flaky testsComing soon

   Tests that pass and fail at random, and how to pin them down.

   15 min readBuilds on: pytest basics (Python)

4. 4. A 40-second report, duplicate rows, dashboards that disagreeComing soon

   Slow reports, duplicate rows and numbers that don't match.

   25 min readBuilds on: Joins: inner, left, right, full, self and cross (SQL) · GROUP BY and HAVING (SQL)

5. 5. A migration must run with zero downtimeComing soon

   Change a busy database's tables without taking the site down.

   20 min readBuilds on: ORMs and migrations (Frameworks)

Advanced

1. 1. Async code runs one call at a time; threads don't speed up CPU workComing soon

   Async code that isn't really running in parallel, and threads that don't speed up maths.

   20 min readBuilds on: Threading vs multiprocessing vs asyncio (Python)

2. 2. Framework incidents: 300 queries on a Django page, a stalled FastAPI endpoint, Flask fails in production, random logoutsComing soon

   Real Django, FastAPI and Flask problems in production.

   30 min readBuilds on: Django ORM performance: N+1, select_related, prefetch_related (Frameworks) · FastAPI async def vs def: the event loop and the threadpool (Frameworks)

3. 3. AI incidents: wrong RAG answers, tripled costs, slow responses, prompt injection, broken JSONComing soon

   When an AI feature gives wrong answers, costs too much or gets tricked.

   30 min readBuilds on: RAG basics: why, chunking, embedding, top-k (AI Engineering)

4. 4. Trade-off questions: SQL or NoSQL for this?Asked oftenComing soon

   Questions with no single right answer: weigh the options out loud.

   20 min readBuilds on: SQL vs NoSQL (System Design)

5. 5. Spot the bugs in a pull request; refactor legacy code safelyComing soon

   Review someone's code, and change old code without breaking it.

   25 min readBuilds on: pytest basics (Python)

Interview-ready

1. 1. Mixed scenario drills, timed and out loudComing soon

   Random scenarios against the clock, answered out loud.

   30 min readBuilds on: Production: memory leak, DB CPU at 100%, deadlocks, disk full, intermittent 502s · Trade-off questions: SQL or NoSQL for this?