diff --git a/Machine Learning/AttendanceSystem.py b/Machine Learning/AttendanceSystem.py
new file mode 100644
index 0000000..355a519
--- /dev/null
+++ b/Machine Learning/AttendanceSystem.py	
@@ -0,0 +1,65 @@
+import cv2
+import numpy as np
+import face_recognition
+import os
+from datetime import datetime
+
+# Step 1: Load known faces
+path = 'faces'
+images = []
+names = []
+
+for file in os.listdir(path):
+    cur_img = cv2.imread(f'{path}/{file}')
+    images.append(cur_img)
+    names.append(os.path.splitext(file)[0])
+
+def encode_faces(images):
+    encode_list = []
+    for img in images:
+        img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
+        encode = face_recognition.face_encodings(img)[0]
+        encode_list.append(encode)
+    return encode_list
+
+print("Encoding known faces...")
+encoded_faces = encode_faces(images)
+print("Encoding complete.")
+
+# Step 2: Mark attendance
+def mark_attendance(name):
+    with open('attendance.csv', 'a+') as f:
+        f.seek(0)
+        lines = f.readlines()
+        if not any(name in line for line in lines):
+            now = datetime.now()
+            dt_string = now.strftime('%Y-%m-%d %H:%M:%S')
+            f.write(f'{name},{dt_string}\n')
+
+# Step 3: Capture live video
+cap = cv2.VideoCapture(0)
+
+while True:
+    success, frame = cap.read()
+    faces_cur_frame = face_recognition.face_locations(frame)
+    encodes_cur_frame = face_recognition.face_encodings(frame, faces_cur_frame)
+
+    for encode_face, face_loc in zip(encodes_cur_frame, faces_cur_frame):
+        matches = face_recognition.compare_faces(encoded_faces, encode_face)
+        face_dis = face_recognition.face_distance(encoded_faces, encode_face)
+        match_index = np.argmin(face_dis)
+
+        if matches[match_index]:
+            name = names[match_index].upper()
+            y1, x2, y2, x1 = face_loc
+            cv2.rectangle(frame, (x1, y1), (x2, y2), (0, 255, 0), 2)
+            cv2.putText(frame, name, (x1, y1 - 10),
+                        cv2.FONT_HERSHEY_SIMPLEX, 0.8, (0, 255, 0), 2)
+            mark_attendance(name)
+
+    cv2.imshow('Face Recognition Attendance', frame)
+    if cv2.waitKey(1) & 0xFF == ord('q'):
+        break
+
+cap.release()
+cv2.destroyAllWindows()