diff --git a/Jenkinsfile b/Jenkinsfile
index ef3df31..3d8e529 100644
--- a/Jenkinsfile
+++ b/Jenkinsfile
@@ -28,8 +28,11 @@ pipeline {
   stages {
     stage('Run create-dataset script') {
       steps {
+        sh 'kaggle datasets download -d mlg-ulb/creditcardfraud'
+        sh 'unzip creditcardfraud.zip'
+        sh 'rm creditcardfraud.zip'
         sh 'chmod +x create-dataset.py'
-        sh 'python3 ./create-dataset.py'
+        sh 'python3 ./create-dataset.py $KAGGLE_USERNAME'
       }
     }
     stage('Archive Artifacts') {
diff --git a/create-dataset.py b/create-dataset.py
index c5e7d0c..46281c3 100644
--- a/create-dataset.py
+++ b/create-dataset.py
@@ -10,9 +10,11 @@ from sklearn.model_selection import train_test_split
 
 
 def download_kaggle_dataset():
-    os.system("kaggle datasets download -d mlg-ulb/creditcardfraud")
-    os.system("unzip creditcardfraud.zip")
-    os.system("rm creditcardfraud.zip")
+    os.environ["KAGGLE_USERNAME"] = "vskyper"
+    os.environ["KAGGLE_KEY"] = sys.argv[1]
+    kaggle = KaggleApi()
+    kaggle.authenticate()
+    kaggle.dataset_download_files("mlg-ulb/creditcardfraud", path="./", unzip=True)
 
 
 def load_data(name):
@@ -96,7 +98,7 @@ def save_whole_data(df, X_train, X_test, y_train, y_test):
 
 
 def main():
-    download_kaggle_dataset()
+    # download_kaggle_dataset()
     os.makedirs("data", exist_ok=True)
 
     df = load_data("creditcard.csv")