diff --git a/.circleci/config.yml b/.circleci/config.yml
index 1ef8c0e33..08bbad6fa 100644
--- a/.circleci/config.yml
+++ b/.circleci/config.yml
@@ -198,6 +198,7 @@ jobs:
               -e AWS_ACCESS_KEY_ID=$AWS_ACCESS_KEY_ID \
               -e AWS_SECRET_ACCESS_KEY=$AWS_SECRET_ACCESS_KEY \
               -e AWS_REGION_NAME=$AWS_REGION_NAME \
+              -e AUTO_INFER_REGION="True" \
               -e OPENAI_API_KEY=$OPENAI_API_KEY \
               -e LANGFUSE_PROJECT1_PUBLIC=$LANGFUSE_PROJECT1_PUBLIC \
               -e LANGFUSE_PROJECT2_PUBLIC=$LANGFUSE_PROJECT2_PUBLIC \
diff --git a/litellm/router.py b/litellm/router.py
index 32c2b61d1..e8b0f658f 100644
--- a/litellm/router.py
+++ b/litellm/router.py
@@ -2559,7 +2559,7 @@ class Router:
 
         # set region (if azure model)
         _auto_infer_region = os.environ.get("AUTO_INFER_REGION", False)
-        if _auto_infer_region == True:
+        if _auto_infer_region == True or _auto_infer_region == "True":
             """
             Hiding behind a feature flag
             When there is a large amount of LLM deployments this makes startup times blow up