Deployment

RAISE is deployed directly from AWS Marketplace using AWS CloudFormation. No source code checkout, Node.js, or CDK toolchain is required.

1. Enable Bedrock Model Access (prerequisite)

Before subscribing, enable model access for global.anthropic.claude-sonnet-4-6 in the target region — open the Bedrock console's Model access page and request access if it isn't already granted.

Required before launch

This is a one-time, per-account, per-region console opt-in — CloudFormation cannot grant it for you. Skipping this step does not fail the stack creation; instead every resume you upload will fail processing silently and land in the dead-letter queue. If CloudWatch shows no Lambda invocations succeeding after a test upload, check this first.

2. Subscribe

Open the RAISE listing on AWS Marketplace and click Continue to Subscribe. Review and accept the terms, then click Continue to Configuration.

3. Choose a Region

Select the AWS Region where RAISE will run.

Bedrock availability

Bedrock is not available in every AWS region. Confirm that your chosen region supports global.anthropic.claude-sonnet-4-6 before proceeding — see the AWS Service Availability page.

4. Launch the CloudFormation Stack

Under Launch this software, select Launch CloudFormation and click Launch.

The AWS CloudFormation console opens with the RAISE template pre-loaded. On the Specify stack details page, provide:

Parameter	Description
Stack name	A name for this deployment — e.g. raise
AlertEmail	Email address for CloudWatch alarm notifications
OutputTarget	Where extracted profiles are sent: DYNAMODB, SQS, or BOTH (default: BOTH)

info

AWS sends a subscription confirmation email to the address you provide. You must confirm it before alarm notifications are delivered.

See Configure Output Target for a full explanation of each option.

Click Next through Configure stack options (no changes required), then Submit.

CloudFormation provisions all RAISE resources — S3 bucket, Lambda function, SQS queues, DynamoDB table, CloudWatch alarms, SNS topic, and dashboard. Deployment typically completes in 2–3 minutes.

5. Stack Outputs

Once the stack status shows CREATE_COMPLETE, open the Outputs tab in the CloudFormation console to find the key resource identifiers:

Output	Description
BucketName	S3 bucket — upload resumes here
TableName	DynamoDB table — stores extracted candidate profiles
OutputQueueUrl	SQS output queue — downstream consumers read from here

Note these values — you will need the bucket name to upload resumes and the queue URL to consume extracted data.

6. Verify the Deployment

Upload a test resume to confirm the end-to-end pipeline is working:

aws s3 cp ./your-resume.pdf s3://YOUR_BUCKET_NAME/resumes/your-resume.pdf

Then check:

1. CloudWatch Dashboard — navigate to RAISE-CV-Ingestion and confirm a Lambda invocation appears with no errors.
2. DynamoDB — open CVExtractorTable and verify a new item exists with the extracted candidate profile.
3. SQS Output Queue — open CVExtractorQueue in the SQS Console and confirm a message is available (if OUTPUT_TARGET is SQS or BOTH).

Deploying to a Different Region

To run RAISE in an additional region, repeat the Marketplace launch from Step 3 and choose the target region. Each region gets its own independent CloudFormation stack.

Ensure Bedrock model access for global.anthropic.claude-sonnet-4-6 is enabled in the target region before launching.

Tearing Down

To remove all RAISE resources from your account, delete the CloudFormation stack.

Console: Open the CloudFormation console, select the RAISE stack, and click Delete.

AWS CLI:

aws cloudformation delete-stack --stack-name raise

Most resources are removed on stack deletion. The resume bucket (BucketName) and the candidate profile table (TableName) are retained — deleting the stack does not delete your candidate data. To permanently remove them, delete the bucket/table manually after the stack delete completes.