GO TO A SECTION

↑ ↓ to navigate · Enter to openExplore all sections →
FDEInterviews logo
← LLM & GenAI Fundamentals / 57
hardAnthropicOpenAI

Here are 30 examples where our prompt gets the wrong answer. Improve it, show me the eval before and after, and do not overfit to the 30.

This is the live practical several labs run, and it is scored on protocol more than on the prompt you end with. Read all thirty before editing, cluster them, hold some out, change one thing at a time, and report with intervals, because 24 of 30 fixed is somewhere between 63% and 91%.

Updated Sep 2026 · Grounded in real Forward Deployed Engineer interview loops and written to a senior-engineer editorial bar.

This is the live practical several labs run, and it is scored on protocol more than on the prompt you end with. Read all thirty before editing, cluster them, hold some out, change one thing at a time, and report with intervals, because 24 of 30 fixed is somewhere between 63% and 91%.

20 answers per topic instead of 10, plus saved progress and bookmarks · no cardor unlock all 545 remaining answers · ₹2,000 / $25

Every answer, concept and course, every paper breakdown, all published video lessons, hands-on FDE Lab missions, Premium PDF guides and companion files, the full practice-test bank and work-sample downloads. Referral Premium excludes guide PDFs and their companion files.

6 months · One payment · No auto-renewal

Study alongside free video lessons.

Keep practising

Next in this trackYour voice agent talks over callers and mishears account numbers, yet word error rate is 6%. Diagnose both, and say what to measure instead.PremiumNext in this trackDesign the eval for a voice agent taking inbound support calls. What do you measure offline, and what can only production tell you?PremiumNext in this trackA customer wants to replace their LLM ticket classifier with Jev to cut cost. What changes, and what do you verify before switching?Premium
FEDITOR'S NOTE

The candidates who do well treat the thirty as symptoms and the eval as the deliverable. The ones who do badly open the prompt in the first minute, fix example one, and by minute forty have a prompt that contains most of the thirty examples verbatim and a held-out score that went down. The interviewer is watching for the split, the regression set, and whether the final report admits what did not get fixed.

DISCUSSION · 0

No comments yet — be the first to share your approach.