Skip to content
mentorship

Math and statistics screen

Timed ML math oral

A correct last line is not enough. The signal is assumptions, key identities, sanity checks, and what the result says about model behavior.

Eight separate drills, 8 to 15 minutes eachPrompt deck, observer follow-up, scored oral derivation

Do this only if the format matches

Confirm the round with recruiting first. This is an original practice exercise, not a claim that a specific company uses this exact prompt.

Protocol

  1. Draw one prompt without previewing its checks.
  2. Define symbols, dimensions, and assumptions aloud.
  3. Derive without notes while the observer interrupts one skipped step.
  4. Check dimensions and one boundary or special case.
  5. Interpret which term changes model behavior and why.
  6. Answer one follow-up that changes an assumption.

Deliverables

  • A recorded or observed derivation.
  • A 0 to 2 score for six dimensions.
  • One corrected derivation written from memory.
  • A spaced retry on another day.

Graduation gates

01

No zero in setup or derivation.

02

At least 9 of 12 scorecard points.

03

A changed-assumption follow-up is handled cleanly.

04

A second spaced success is recorded before graduation.

Starter files

Prompt deckEight derivations with observer checks.Open file →ScorecardSix dimensions and a passing gate.Open file →

Answer key is a rubric, not a solution dump

The linked question explains what strong performance looks like, common failure modes, and senior follow-ups. Attempt the lab before reading it.

Read the oral-screen strategy →

← All executable labs