Ruby + Go + AWS + Postgres
Architecture cert — "which services solve this problem." No code, multiple choice. This is the foundation. Every other cert and project builds on this vocabulary.
- Fireship "AWS in 10 Minutes" — watch this one first, fast orientation
- Adrian Cantrill free YouTube explainers — pick services you're curious about (S3, RDS, Lambda, VPC)
- freeCodeCamp full SA Associate course — listenable on walks
- TechWorld with Nana, Cloud Resume Challenge podcast
More search terms: AWS SA Associate 2025, AWS services explained, AWS S3 explained, AWS Lambda explained, AWS VPC explained, AWS well architected framework, AWS re:Invent keynote 2024
When the videos start feeling familiar instead of new, you're ready for the course.
- Adrian Cantrill SA Associate course ($40) — structured, much listenable without a screen
- Tutorial Dojo practice exams ($15) — closest thing to the real exam
Pick these up when a video mentions them. All free at learning scale.
| Tech | What to try | CLI |
|---|---|---|
| S3 | Create a bucket, upload a file, host a static page | aws s3 mb, aws s3 cp |
| IAM | Create a user, attach a policy, use access keys | aws iam |
| Lambda | Function that processes an S3 upload | aws lambda or CDK |
| API Gateway | HTTP endpoint in front of a Lambda | CDK or SAM |
| DynamoDB | Create a table, put/get items | aws dynamodb |
| SQS | Send messages between two Lambdas | aws sqs |
| CloudWatch | Look at logs from Lambda runs | aws logs |
Learn Docker locally (free) — Dockerfiles, docker-compose, images, networking. That's the transferable skill. ECS/Fargate is just "run this container on AWS" and costs ~$10-30/month, so spin up once to prove it, then tear down. The exam only needs conceptual understanding.
YouTube: Docker tutorial beginner, Dockerfile explained, docker compose tutorial, ECS vs EKS vs Fargate
You know Postgres. RDS is Postgres managed by AWS — same SQL, same psql, same pg gem. The exam tests when to pick RDS vs Aurora vs DynamoDB, not SQL.
The cheat: Use Neon (neon.tech, free tier) for practice. Same Postgres, accessible over the internet. Save actual RDS for a single prove-it-works afternoon (~$0.50).
YouTube: AWS RDS vs Aurora vs DynamoDB, when to use DynamoDB vs RDS
- AWS CLI —
aws <service> <command>. Start here to understand what each service does. - CDK — Infrastructure as code in TypeScript.
cdk deploy. The "text editor + deploy" workflow. - SAM — Lighter than CDK, Lambda/API Gateway focused. YAML +
sam deploy.
YouTube: AWS CLI tutorial beginner, AWS CDK tutorial 2025, Infrastructure as Code explained, AWS CDK vs Terraform
Using AI services (Bedrock), not training models. Calling APIs, RAG patterns. Build a small RAG project — covers the cert material and gives you a real project for your resume. Bedrock costs ~$1-2/month at learning scale.
Concepts: RAG, embeddings, vector stores, prompt engineering, responsible AI, model selection, fine-tuning vs prompting, tokens, Bedrock vs SageMaker (Bedrock = call API, SageMaker = train/host model)
YouTube: AWS Bedrock tutorial, RAG explained simply, AWS AI Practitioner exam 2025, vector database explained, LLM tokens explained, building AI apps AWS
Study: AWS Skill Builder (free) has AI Practitioner prep. Tutorial Dojo practice exams.
Data pipelines, analytics, ETL. This gives you the vocabulary — data lakes, star schemas, Redshift, Glue, Kinesis — for concepts you already understand but can't name-drop in interviews.
Concepts: ETL/ELT, data lakes vs data warehouses, star schema, fact/dimension tables, streaming vs batch, data catalog, partitioning, columnar storage
YouTube: data lake vs data warehouse explained, star schema dimensional modeling, AWS Glue tutorial, AWS Kinesis explained, AWS Redshift explained, ETL pipeline explained, AWS Data Engineer Associate 2025
Study: Stephane Maarek on Udemy (~$15 on sale). Tutorial Dojo practice exams.
Lambda, DynamoDB patterns, API Gateway, SQS, Step Functions, CI/CD. Has hands-on code questions. By this point you'll have touched most of these from earlier goals.
YouTube: AWS Developer Associate 2025, AWS Step Functions explained, DynamoDB single table design, AWS CodePipeline tutorial, AWS X-Ray tracing, serverless architecture patterns
Study: Stephane Maarek on Udemy (~$15 on sale). Tutorial Dojo practice exams.
| Tech | What it does | Cost |
|---|---|---|
| Step Functions | Orchestrate multi-step workflows | Free tier: 4,000 transitions/month |
| DynamoDB Streams | React to database changes | Free with DynamoDB |
| CodePipeline | CI/CD — build, test, deploy | 1 free pipeline, then $1/month |
| X-Ray | Trace requests across services | Free tier: 100K traces/month |
| Cognito | User auth/signup | Free tier: 50K monthly active users |
| EventBridge | Event routing between services | Free for AWS events |
| Technology | AWS name | Transfers to | Notes |
|---|---|---|---|
| Object storage | S3 | GCS, Azure Blob | S3 is the standard |
| Serverless functions | Lambda | Cloud Functions, Azure Functions | Free to learn |
| Managed Postgres | RDS / Aurora | Any managed Postgres | Use Neon for free practice |
| NoSQL | DynamoDB | MongoDB, etc. | Free tier is generous |
| Containers | ECS / Fargate | Kubernetes anywhere | Docker locally (free) |
| Message queues | SQS | RabbitMQ, Kafka | Free to learn |
| Infra as code | CDK | Terraform, Pulumi | CDK uses TypeScript — bonus TS exposure |
| AI/ML APIs | Bedrock | OpenAI API, Azure AI | Pennies at learning scale |
| Data pipelines | Glue / Kinesis | Spark, Kafka, Airflow | Vocabulary for interviews |
- Free: Lambda, DynamoDB, SQS, S3, IAM, CloudWatch, API Gateway, Step Functions, Bedrock (pennies)
- Expensive: RDS (
$15/month), ECS/Fargate ($10-30/month), EC2 - The gotcha: Leaving an RDS or EC2 instance running 24/7. Set a billing alarm for $10:
aws cloudwatch put-metric-alarm
Strategy: Use free services for learning. Use Neon for Postgres. Use Docker locally for containers. Spin up RDS/ECS only to prove it works, then tear down.