Glean's signature design question. The two traps: post-filtering that drains your top-k, and the leak through generated answers. The winning architecture filters at retrieval, and here is how, at scale.
more free answers with an account · no card
← RAG & Agent System Design / 20
hardGleanMicrosoftOpenAI
Design RAG over 10M documents with per-user permissions: a user must never see data they can't access
Glean's signature design question. The two traps: post-filtering that drains your top-k, and the leak through generated answers. The winning architecture filters at retrieval, and here is how, at scale.
Updated Aug 2026 · Grounded in real Forward Deployed Engineer interview loops and written to a senior-engineer editorial bar.
UP NEXT ON YOUR JOURNEY
Learners often go here nextGive an agent memory across sessions without blowing up the context window or leaking between usersNext in this trackYour agent fails 5% of the time mid-workflow, after executing real side effects. Make it safe.Next in this trackDesign an AI contract-review workflow (clause extraction, risk flags, citations) that skeptical lawyers will actually trust
DISCUSSION · 0
No comments yet — be the first to share your approach.