diff --git a/_questions/machine-learning-zoomcamp/module-2/033_2885d25af1_pandas-fillna-missing-values-not-updated.md b/_questions/machine-learning-zoomcamp/module-2/033_2885d25af1_pandas-fillna-missing-values-not-updated.md
new file mode 100644
index 00000000..cef04c85
--- /dev/null
+++ b/_questions/machine-learning-zoomcamp/module-2/033_2885d25af1_pandas-fillna-missing-values-not-updated.md
@@ -0,0 +1,30 @@
+---
+id: 2885d25af1
+question: Why are there still missing values after I use `fillna()` in Pandas?
+sort_order: 33
+---
+
+By default, Pandas operations like `fillna()` don’t modify the original DataFrame in-place; they return a modified copy. If you don’t assign that result back, the missing values remain.
+
+Fix: assign the returned Series/column back, e.g.:
+- `df['score'] = df['score'].fillna(fill_value)`
+
+Alternative: you can write into the existing data by using `inplace=True`, e.g.:
+- `df['score'].fillna(fill_value, inplace=True)`
+
+Example:
+```python
+import pandas as pd
+import numpy as np
+
+df = pd.DataFrame({"score": [100, np.nan, 150]})
+fill_value = 0
+
+# INCORRECT: change is lost because the result isn't assigned
+df["score"].fillna(fill_value)
+print(df["score"].isnull().sum())  # 1
+
+# CORRECT: reassign the column
+df["score"] = df["score"].fillna(fill_value)
+print(df["score"].isnull().sum())  # 0
+```
\ No newline at end of file