Start with a closed-book baseline
Do not read the full path first.
Attempt the first scored question now. Use the result to decide which later steps need repair. Start first practice- 01
Calibrate role and level
Read the role taxonomy and senior-through-principal calibration; choose the target rubric.
- 02
Scope before solving
Practice objective, user, data, constraints, and the smallest discriminating experiment.
- 03
Prove ML breadth
Use distributions, costs, mechanisms, and alternatives rather than memorized recipes.
- 04
Design end to end
Connect data, baseline, model, evaluation, serving, monitoring, and iteration.
- 05
Make a product decision
Define the causal estimand, metrics, validity checks, uncertainty, and ship rule.
- 06
Debug evidence disagreement
Separate objective mismatch, point-in-time leakage, serving skew, and feedback effects.
- 07
Defend project ownership
Prepare a short opening and a 20-minute technical deep dive.
- 08
Show judgment under pressure
Balance legitimate delivery pressure with measurable user risk.
- 09
Simulate the loop
Run the AS packet and reduce results to three bounded repairs.
After the core path
Add only the format and subject that recruiting confirmed. Do not add every frontier lab or specialist topic.
Choose one domain supplement · Add the upper-IC level overlay · Check current lab formats · Run the role simulation