Implementing Exponential Backoff for Genesys Cloud /api/v2/users Bulk Updates

Can anyone clarify the correct backoff implementation for bulk user updates? I am hitting 429 errors with this Python snippet:

for user in users:
 response = requests.put(url, json=data)
 if response.status_code == 429:
 # Immediate retry fails
 pass

1. Loop through 50 users
2. PUT /api/v2/users/{id}
3. Receive 429

What is the recommended delay strategy? Should I parse the Retry-After header or use a fixed exponential backoff?