AAI Agentic Handbook
Staff track

Owning the trade-offs

Org-wide platforms, security and saying no well — and how to make that case in a design review or an interview.

Staff · Owning the trade-offs
  1. Read · 6 minWhat AI-engineering interviewers test at each levelOne design question, three scored answers (Mid, Senior, Staff), the rubric behind the scores, and the red flags that sink otherwise good candidates.
  2. Learn a term · quiz · ~8 minTrade-off framingChoosing one option knowing what it costs you on another axis. Senior answers name the axes, the deciding constraint, and when they'd switch.
  3. Learn a term · quiz · ~8 minStaff engineer scopeA senior IC level whose scope spans teams: sets technical direction, de-risks big bets, and multiplies others, mostly through writing and influence.
  4. Read · 6 minTalking a team out of RAG (when it is the wrong tool)RAG is a default, not a requirement. When long context with prompt caching or text-to-SQL is the better design, what it costs and how fast it is with numbers, how per-document permissions change the answer, and the written threshold that flips the decision.
  5. Interview question · ~15 minHow do you decide whether a problem actually needs RAG?
  6. Learn a term · quiz · ~8 minMulti-tenancyServing many customers from shared infrastructure while keeping their data, quotas and performance isolated. Isolation tier is a cost vs risk choice.
  7. Interview question · ~15 minHow would you prevent one customer's documents from appearing in another customer's results?
  8. Read · 8 minSecuring agents: assume the model will be fooledPrompt injection has no reliable filter, so agent security is architecture: break the lethal trifecta, constrain what a fooled model can do, and test it like any other regression.
  9. Learn a term · quiz · ~8 minPrompt injection (indirect)Text that the model reads as instructions: typed by a user (direct) or hidden in a page, email or doc the agent reads (indirect). No filter fully fixes it.
  10. Learn a term · quiz · ~8 minBlast radiusHow much breaks when one thing fails or is compromised: how many users, tenants, records or actions. Design to keep it small and known.
  11. Learn a term · quiz · ~8 minSTAR methodSituation, Task, Action, Result: a structure for behavioural interview answers. Spend most of the time on what you did and what changed, with numbers.
  12. Interview question · ~15 minWhat would you do if RAG quality suddenly drops after a production deployment?