Databricks-ML-Assoc Databricks Machine Learning Practice Question
A data scientist is using Databricks Feature Store to build training sets and wants to ensure the features used at training time are consistent with those served at inference time. Which TWO practices help guarantee this consistency? (Choose two.)
⚠ Common exam trap
The trap here is believing that reimplementing features for serving speed is acceptable, when it actually creates train/serve skew.
Answer choices
Why each option matters
Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.
Correct answer & explanation
✓
Publish the same feature computations into the online store so low-latency serving reads identical values.
Consistency comes from reusing the same feature definitions and lookups across training and serving. Creating a training set with create_training_set preserves lineage so models know which features to retrieve, and publishing the same computations to the online store ensures low-latency serving reads values produced by identical logic, eliminating train/serve skew.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Disable feature lineage tracking to reduce metadata overhead during training.
Why it's wrong here
Disabling lineage removes the metadata that links a model to the features and lookups it used, making it impossible for the serving path to reconstruct the training feature set automatically. That metadata is precisely what enables consistent retrieval. Turning it off increases the risk of mismatched features and breaks the automated consistency guarantees the Feature Store offers.
- ✗
Recompute features with different SQL logic at serving time to optimize latency.
Why it's wrong here
Using different computation logic at serving time than at training time is a classic source of training/serving skew. Even if the new logic is faster, it can produce different values for the same entity, so the model receives inputs unlike those it learned from. Feature Store consistency depends on reusing the same definitions rather than rewriting them for serving.
- ✓
Publish the same feature computations into the online store so low-latency serving reads identical values.
Why this is correct
The online store holds materialized feature values for low-latency lookups and is populated from the same feature tables used for training. Publishing to the online store ensures the inference path retrieves the values defined by the same computation, preventing drift between offline training features and online serving features, which is the core consistency guarantee the Feature Store provides.
- ✗
Copy feature values into a CSV file and load it in the serving notebook.
Why it's wrong here
Manually copying features into a CSV disconnects the serving path from the feature tables and their update schedule. Values can become stale, and the transformation logic is no longer guaranteed to match training. This manual approach undermines consistency because it bypasses the Feature Store's managed lookups and lineage, reintroducing the exact train/serve skew the team wants to avoid.
- ✓
Create a training set with create_training_set so the model records feature lineage and lookups.
Why this is correct
Using create_training_set ties the training data to the feature tables and produces metadata that a model can use to retrieve the same features at inference. This lineage lets the scoring path automatically join the correct feature values, which directly supports train/serve consistency because the same feature definitions and lookups are reused rather than reimplemented manually.
About these practice questions
Courseiva writes every Databricks-ML-Assoc question from scratch — 319 in total, each with an explanation and a wrong-answer breakdown. None are copied from real exams or dumps. Learn why practice questions differ from exam dumps →
JA
Written and reviewed by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
Last reviewed September 2026 · checked against the official Databricks exam blueprint
This Databricks-ML-Assoc practice question is part of Courseiva's free Databricks certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the Databricks-ML-Assoc exam.