Ai Coding And Product Systems medium
Your workspace AI filters retrieval using ACLs stored in the vector index, which is rebuilt nightly. An admin removes an employee from a confidential project page, but for the rest of the day the employee can still get AI answers drawn from it. What is the right fix?
- A Check retrieved chunks against the permission source of truth at query time, and push ACL changes to the index via events.
- B Shorten the nightly re-index to an hourly job, since a one-hour exposure window is acceptable for most enterprise permission changes.
- C Instruct the model to ignore pages whose metadata shows the user was recently removed, using a timestamp passed in the prompt.
- D Cache each user's permission set in the session for its full lifetime so retrieval avoids repeated calls to the permissions service.
Answer: A. Check retrieved chunks against the permission source of truth at query time, and push ACL changes to the index via events.
A query-time authorization check plus **event-driven ACL updates** trades a little latency (and some lost recall when many chunks get filtered out) for revocations that take effect almost immediately. Hourly reindexing (B) is tempting because it is cheap, but it still leaks revoked content, and the model cannot enforce permissions (C). Batch ACL sync is acceptable only for low-sensitivity, mostly public workspaces.
One system design question a day, with spaced repetition so the ones you miss come back.
Get the appRelated questions
- Your assistant uses a classifier to route queries to a small or large model, which saves 60% of GPU spend. Analysis shows about 20% of hard queries are misrouted to the small model, which then gives confident wrong answers, and regenerate clicks on those threads have doubled. What should you do?
- Your workspace AI retrieves the top-k chunks from a shared vector index and passes them to the model with ACL metadata and a system prompt saying 'only use documents this user can access.' A user just received an answer quoting a private HR page they cannot open. What is the correct fix?
- Your support bot answers billing questions. One user has asked the same refund question three times, and retrieval confidence for that topic is low. What should the bot do next?
- Your AI review system receives a 4,000-line pull request spanning 60 files. Sending the whole diff in one prompt yields shallow, generic comments, and real bugs in files near the middle of the diff are missed. How should you structure the review?
- In your ChatGPT-style assistant, power users keep single threads running for weeks, reaching about 150K tokens. Cost per turn keeps climbing, latency is rising, and the model increasingly misses facts from early in the thread. What should you do?
- Free users consume 70% of your assistant's GPU hours but bring in about 5% of revenue through conversions. Leadership wants to keep the free tier useful enough to drive upgrades. Which policy fits?