diff --git a/.github/workflows/gpt-to-redis-batch-job.yml b/.github/workflows/gpt-to-redis-batch-job.yml
new file mode 100644
index 0000000..4b4f26d
--- /dev/null
+++ b/.github/workflows/gpt-to-redis-batch-job.yml
@@ -0,0 +1,34 @@
+name: GPT To Redis Batch Job
+
+on:
+  workflow_dispatch:
+  push:
+    branches:
+      - main
+  schedule:
+    - cron: '0 2 * * *'
+
+jobs:
+  batch-job:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v3
+      - name: Docker Build And Push
+        run: |
+          docker login -u ${{ secrets.USERNAME }} -p ${{ secrets.PASSWORD }}
+          docker build -f gpt.Dockerfile -t skku-gpt-to-redis-batch .
+          docker tag skku-gpt-to-redis-batch sinkyoungdeok/skku-gpt-to-redis-batch
+          docker push sinkyoungdeok/skku-gpt-to-redis-batch
+      - name: Deploy Prod
+        uses: appleboy/ssh-action@v0.1.4
+        with:
+          key: ${{ secrets.SSH_KEY }}
+          host: ${{ secrets.HOST_NAME }}
+          username: ubuntu
+          port: 22
+          script: |
+            docker login -u ${{ secrets.USERNAME }} -p ${{ secrets.PASSWORD }}
+            docker pull sinkyoungdeok/skku-gpt-to-redis-batch
+            
+            docker run --net ubuntu_default -e OPENAI_API_KEY=${{ secrets.OPENAI_API_KEY }} sinkyoungdeok/skku-gpt-to-redis-batch
+      
diff --git a/csv-to-gpt.py b/csv-to-gpt.py
new file mode 100644
index 0000000..1609ffa
--- /dev/null
+++ b/csv-to-gpt.py
@@ -0,0 +1,128 @@
+import json
+import os
+
+import openai
+import pymysql
+import redis
+
+# MySQL 연결 설정
+conn = pymysql.connect(
+  host="skku-db",
+  port=3306,
+  user="skku-user",
+  password="skku-pw",
+  db="skku",
+  charset='utf8'
+)
+
+# Redis 연결 설정
+redis_client = redis.StrictRedis(host='skku-redis', port=6379, db=0,
+                                 decode_responses=True)
+
+SEARCH_PREFIX = "SR:"
+MAX_HISTORY = 5
+
+# OpenAI API 키 설정
+openai.api_key = os.getenv("OPENAI_API_KEY")
+
+
+def get_search_queries(user_id):
+  key = f"{SEARCH_PREFIX}{user_id}"
+  return redis_client.lrange(key, 0, MAX_HISTORY - 1)
+
+
+def get_filtered_restaurants():
+  query_restaurants = """
+    SELECT id, name, original_categories, naver_rating_avg, naver_review_count
+    FROM restaurants
+    WHERE naver_rating_avg >= 4.5 AND naver_review_count >= 200;
+    """
+  with conn.cursor() as cursor:
+    cursor.execute(query_restaurants)
+    return cursor.fetchall()
+
+
+def get_user_liked_restaurants():
+  query_liked_restaurants = """
+    SELECT u.id AS user_id, r.name AS restaurant_name, r.original_categories AS category
+    FROM users u
+    JOIN restaurant_likes rl ON u.id = rl.user_id
+    JOIN restaurants r ON rl.restaurant_id = r.id;
+    """
+  with conn.cursor() as cursor:
+    cursor.execute(query_liked_restaurants)
+    return cursor.fetchall()
+
+
+def get_gpt_recommendations(user_data, restaurant_data):
+  prompt = f"""
+    유저가 데이터랑, 음식점 데이터를 보여줄게.
+    이걸 기반으로 음식점을 20개 추천해줘.
+    다른 이야기는 다 빼고 추천할 음식점id를 ","기반으로 구분해서 리스트로 알려줘
+
+    User Data:
+    {json.dumps(user_data)}
+
+    Restaurant Data:
+    {json.dumps(restaurant_data)}
+    """
+  response = openai.ChatCompletion.create(
+    model="gpt-4-turbo",
+    messages=[
+      {"role": "system", "content": "You are a helpful assistant."},
+      {"role": "user", "content": prompt}
+    ],
+    max_tokens=2000
+  )
+  return response['choices'][0]['message']['content'].strip()
+
+
+def save_recommendations_to_redis(user_id, recommendations):
+  key = f"RECOMMENDATION:{user_id}"
+  redis_client.delete(key)
+  for recommendation in recommendations.split('\n'):
+    redis_client.rpush(key, recommendation)
+  redis_client.expire(key, 3600 * 24 * 3)
+
+def save_all_restaurants_to_redis(filtered_restaurants):
+  all_restaurant_ids = [str(restaurant[0]) for restaurant in filtered_restaurants]
+  recommendations_string = ','.join(all_restaurant_ids)
+  save_recommendations_to_redis(0, recommendations_string)
+
+
+filtered_restaurants = get_filtered_restaurants()
+user_liked_restaurants = get_user_liked_restaurants()
+
+user_data = {}
+for liked_restaurant in user_liked_restaurants:
+  user_id = liked_restaurant[0]
+  restaurant_name = liked_restaurant[1]
+  category = liked_restaurant[2]
+
+  if user_id not in user_data:
+    user_data[user_id] = {'liked_restaurants': [], 'recent_searches': []}
+
+  user_data[user_id]['liked_restaurants'].append({
+    'restaurant_name': restaurant_name,
+    'category': category
+  })
+
+for user_id in user_data.keys():
+  recent_searches = get_search_queries(user_id)
+  user_data[user_id]['recent_searches'] = recent_searches
+
+restaurant_data = [dict(zip(
+  ['id', 'name', 'original_categories', 'naver_rating_avg',
+   'naver_review_count'], restaurant)) for restaurant in filtered_restaurants]
+
+print("restaurant_data", restaurant_data, "length:", len(restaurant_data))
+for user_id, data in user_data.items():
+  print("user_id", user_id, "data", data)
+  try:
+    recommendations = get_gpt_recommendations(data, restaurant_data)
+  except:
+    continue
+  print(recommendations)
+  save_recommendations_to_redis(user_id, recommendations)
+
+save_all_restaurants_to_redis(filtered_restaurants)
diff --git a/gpt-requirements.txt b/gpt-requirements.txt
new file mode 100644
index 0000000..0ff49bf
--- /dev/null
+++ b/gpt-requirements.txt
@@ -0,0 +1,3 @@
+openai==0.28
+redis
+pymysql
diff --git a/gpt.Dockerfile b/gpt.Dockerfile
new file mode 100644
index 0000000..cb06516
--- /dev/null
+++ b/gpt.Dockerfile
@@ -0,0 +1,8 @@
+FROM python:3.8-slim
+
+COPY gpt-requirements.txt gpt-requirements.txt
+COPY csv-to-gpt.py csv-to-gpt.py
+
+RUN pip install -r gpt-requirements.txt
+
+ENTRYPOINT ["python3", "csv-to-gpt.py"]
\ No newline at end of file