deploy

name: deploy description: Deploy applications to AWS (SageMaker, Amplify, EC2). Use this skill to deploy models, frontends, or manage infrastructure. Invoke with /deploy.

AWS Deployment

This skill manages deployments to AWS services for the wc_simd project.

SageMaker Endpoints

Deploy Embedding Model

cd demos/timetrvlr/cdk
npm install
cdk deploy

Or manually:

import sagemaker
from sagemaker.huggingface import HuggingFaceModel

model = HuggingFaceModel(
    model_data="s3://bucket/model.tar.gz",
    role="arn:aws:iam::xxx:role/SageMakerRole",
    transformers_version="4.37",
    pytorch_version="2.1",
    py_version="py310"
)

predictor = model.deploy(
    instance_type="ml.g5.xlarge",
    endpoint_name="embedding-endpoint"
)

Async Inference

For long-running inference (VLM embeddings):

from sagemaker.async_inference import AsyncInferenceConfig

async_config = AsyncInferenceConfig(
    output_path="s3://bucket/async-output/",
    max_concurrent_invocations_per_instance=4
)

predictor = model.deploy(
    instance_type="ml.g5.2xlarge",
    async_inference_config=async_config
)

AWS Amplify (Frontend)

TimeTraveler Demo

cd demos/timetrvlr/amplify-cdk
npm install
cdk deploy

The CDK stack:

• Connects to GitHub repository
• Sets up build pipeline
• Configures custom domain (optional)
• Deploys Next.js/React frontend

Manual Amplify Setup

amplify init
amplify add hosting
amplify publish

EC2 Instances

Start/Stop via Script

python aws/ec2_control.py start --name simd_gpu
python aws/ec2_control.py stop --name simd_gpu

Launch New Instance

Use AWS Console or CLI:

aws ec2 run-instances \
  --image-id ami-xxx \
  --instance-type g5.xlarge \
  --key-name your-key \
  --security-group-ids sg-xxx \
  --iam-instance-profile Name=spark-docker-s3-profile

S3 Data Management

Upload Data

aws s3 sync data/ s3://bucket/data/

Download Data

aws s3 sync s3://bucket/data/ data/

RDS (Hive Metastore)

The production Spark stack uses RDS MySQL for the Hive metastore.

Connect Manually

mysql -h <rds-endpoint> -u hive -p hive

Initialize Schema

Set INIT_HIVE_SCHEMA=true in spark_docker_s3/.env on first run.

CDK Stacks

Deploy CDK Stack

cd <stack-directory>
npm install
cdk bootstrap  # First time only
cdk synth      # Preview
cdk deploy     # Deploy

Destroy Stack

cdk destroy

Environment Variables

Required in .env:

AWS_REGION=eu-west-2
S3_BUCKET=your-bucket
HIVE_METASTORE_HOST=rds-endpoint
HIVE_METASTORE_USER=hive
HIVE_METASTORE_PASSWORD=xxx

Load with:

from dotenv import load_dotenv
load_dotenv()