Question bank
Pick an answer to see if you're right — with an explanation.
All topicsPythonSQL / MySQLData ScienceNumPyPandasMatplotlibMachine LearningDeep LearningArtificial IntelligenceLarge Language ModelsDjangoAPIsGit & GitHubPower BIExcelStatistics
Deep LearningEasy
What does `model.eval()` change?
aIt disables gradient computation entirely
bIt freezes all parameters so they cannot be trained
cIt moves the model to the CPU
dIt switches layers like dropout and batch norm to inference behaviour
Explanation: `eval()` changes layer behaviour only; use `torch.no_grad()` separately to stop gradient tracking.
DjangoEasy
Which default permission codename does Django create for editing a `Product` in app `shop`?
ashop.edit_product
bshop.change_product
cshop.update_product
dshop.modify_product
Explanation: Django auto-creates add_, change_, delete_ and view_ permissions for each model.
ExcelEasy
In Data Validation, which Allow option creates a dropdown list in a cell?
aDecimal
bCustom
cList
dText length
Explanation: The List option builds a dropdown from typed values or a range reference.
Git & GitHubEasy
Where do GitHub Actions workflow files live?
aIn the repo root as actions.yml
bIn GitHub account settings only
cIn a .git/hooks folder
dIn the .github/workflows/ directory
Explanation: GitHub discovers YAML workflow files under .github/workflows/.
Large Language ModelsEasy
Which approach is best for answering questions about a price list that changes every week?
aFine-tuning every week
bA longer system prompt with old prices
cHigher temperature
dRAG over the current price list or a tool that queries it
Explanation: Fresh, changing facts belong in retrieval or tools, not in model weights.
Machine LearningEasy
Why must features usually be scaled before k-means?
aK-means only accepts values between 0 and 1
bScaling makes it run on GPUs
cScaling removes outliers
dIt uses distances, so large-unit features would dominate
Explanation: Euclidean distance is dominated by features with large numeric ranges unless they are standardised.
MatplotlibEasy
Why should you avoid using only red vs green to separate profit and loss series?
aMatplotlib cannot draw red
bGreen lines are slower to render
cIt increases file size
dMany readers have red-green colour vision deficiency
Explanation: Colour-blind readers may not distinguish them; add markers, labels or use a safe palette.
NumPyEasy
What is the shape of `np.arange(12).reshape(3, -1)`?
a(3, 3)
b(12, 3)
c(3, 12)
d(3, 4)
Explanation: -1 lets NumPy infer the size: 12 elements / 3 rows = 4 columns.
PandasEasy
Which dtype is best for a column with only 5 distinct payment modes repeated across millions of rows?
aobject
bfloat64
cstring
dcategory
Explanation: category stores each distinct value once plus compact integer codes.
Power BIEasy
Which object is computed row by row and physically stored in the model at refresh time?
aA measure
bA slicer
cA report-level filter
dA calculated column
Explanation: Calculated columns are materialised into the table at refresh; measures are evaluated per visual cell at query time.
PythonEasy
With `from pathlib import Path`, what is `Path('data/sales.csv').stem`?
asales
bsales.csv
c.csv
ddata
Explanation: `.stem` is the file name without its suffix.
SQL / MySQLEasy
In `EXPLAIN` output, what does `type = ALL` indicate?
aAll indexes were used
bA full table scan
cThe query returns all columns
dThe result is cached
Explanation: ALL is the worst access type: MySQL reads every row of the table.
StatisticsEasy
Which test checks whether two categorical variables, such as city and payment method, are related?
aOne-way ANOVA
bChi-square test of independence
cPearson correlation
dPaired t-test
Explanation: Chi-square compares observed counts in a contingency table with counts expected under independence.
PythonEasy
What is `3 // 2`?
a1.5
b2
c0
d1
Explanation: `//` is floor division.
SQL / MySQLMedium
Which is best for storing money in MySQL?
aDECIMAL
bFLOAT
cDOUBLE
dVARCHAR
Explanation: DECIMAL avoids floating-point rounding errors.