This should be faster than looping:

# Using boolean indexing to find the cutoff index
cutoff_index = df[df['stage'] != 0].index.min()

# Slice and reset index
df = df.iloc[cutoff_index:].reset_index(drop=True)
print(df)