AWS aims this exam at people with a year or more of hands-on experience building on AWS, so this plan revises rather than teaches from scratch. It assumes about an hour a day on weekdays and a longer session at the weekend. If you have not yet built anything on Lambda, DynamoDB and API Gateway, do that first and allow six to eight weeks. This exam rewards having seen things fail.
Know the exam first
The AWS Certified Developer - Associate exam (DVA-C03) has 65 questions and lasts 130 minutes. Only 50 questions count towards your score, but the 15 unscored ones are not marked, so treat every question as if it counts. The passing score is 720 on a scale of 1000. It covers four domains:
- Development with AWS Services: 30% of the exam
- Security: 26% of the exam
- Testing and Deployment: 22% of the exam
- Troubleshooting and Optimization: 22% of the exam
Read the official exam guide before you start, including its lists of in-scope and out-of-scope services. The details can change.
What changed from DVA-C02
AWS says DVA-C02 can be taken until November 30, 2026, and DVA-C03 is in use from December 1, 2026. Older courses and notes still cover most of what you need, but check that you have these additions:
- Securing AI services. A new task covers managing access to AI services, keeping data private, filtering model input and output, securing AI agents and protecting prompts in logs.
- AI-assisted development. Using tools to generate, review and test code, and integrating managed AI services into an application.
- Containers. Building and managing images in Amazon ECR, and deploying to Amazon ECS, Amazon EKS and AWS Fargate.
- Presigned URLs. Time-limited access through Amazon S3 and Amazon CloudFront is now called out directly.
Troubleshooting and Optimization also rose to 22% of the exam, so do not leave it until the end. The exam does not expect you to choose foundation models, engineer prompts, design RAG architectures or design pipelines from scratch.
Week 1: Development with AWS Services
- Day 1, Baseline. Take Set 1 under timed conditions, with no looking anything up. Read every explanation and note which topics you need most.
- Day 2, Lambda. Memory, timeout, reserved and provisioned concurrency, layers, VPC access, destinations and dead-letter queues. Know which invocation types retry, and how.
- Day 3, APIs and events. API Gateway proxy and non-proxy integrations, validation and error codes, then SQS, SNS, EventBridge and Step Functions. Practise choosing between synchronous and asynchronous designs.
- Day 4, DynamoDB. Partition keys, query versus scan, LSI versus GSI, consistency, conditional writes, transactions and capacity calculations. Work the arithmetic by hand.
- Day 5, Data and streaming. Kinesis shards and consumers, S3 lifecycle rules and multipart uploads, TTL, and which cache fits which data store.
- Day 6, AI in development. Kiro specs, steering and hooks, then call a model through the Amazon Bedrock Converse API, including tool use and streaming.
- Day 7, Set 2. Take Set 2 and review it. It concentrates on serverless failure modes.
Week 2: Security, Testing and Deployment
- Day 8, Identity. IAM policy evaluation, roles and trust policies, STS, Cognito user pools versus identity pools, and API Gateway authorizers.
- Day 9, Encryption and secrets. Envelope encryption with KMS, SSE-S3, SSE-KMS and SSE-C, ACM Region rules, presigned URLs, Secrets Manager and Parameter Store.
- Day 10, AI security. Controlling model access with IAM, private connectivity through VPC endpoints, Bedrock Guardrails, securing agent tools, and protecting model invocation logs.
- Day 11, Artifacts and containers. Dockerfiles, image tags and digests, ECR lifecycle policies, ECS task definitions, task roles versus execution roles, and EKS deployments.
- Day 12, Testing. SAM local commands, CDK assertions, API Gateway mock integrations, realistic test events, and Lambda aliases and stages as environments.
- Day 13, CI/CD. CodeBuild, CodePipeline and CodeDeploy, AppSpec hooks, CloudFormation change sets, and blue/green, canary and rolling deployments with rollback.
- Day 14, Set 3. Take Set 3, which leans on containers, CI/CD and infrastructure as code.
Week 3: Troubleshooting, Optimization and rehearsal
- Day 15, Observability. CloudWatch Logs Insights queries, structured logging, custom metrics and the embedded metric format, alarms, tracing, and health checks.
- Day 16, Optimization. Lambda memory tuning, connection pooling and RDS Proxy, SQS batching, and caching with ElastiCache, DAX and CloudFront. Read through common error codes and what each one points to.
- Day 17, Confusing pairs. Take Set 4. Write a one-line difference for each pair you miss.
- Day 18, Fix your weakest domain. Use your results pages to find it, then go back to the AWS documentation for those topics.
- Day 19, Scenario wording. Take Set 5, which adds longer scenarios and the AI topics.
- Day 20, Build something. Deploy a small API with SAM, break it on purpose and find the fault in the logs and traces. An afternoon of this makes many troubleshooting questions easy.
- Day 21, Dress rehearsal. Take Set 6, the hardest set, exactly as you would on exam day. If you score 80% or more, you are ready to book. If not, spend a few more days on your weakest domain.
How to get the most from each set
- Pace yourself. 130 minutes for 65 questions is about 120 seconds each. Flag a long scenario and come back to it.
- Read the error first. In troubleshooting questions, the status code or exception name usually narrows the answer down before you read the options.
- Get every part right. On a question that asks you to choose two or three, you need all the correct answers to earn the mark, and there is no penalty for guessing.
- Watch the wording. "Least operational overhead", "most secure" and "without changing the code" often decide between two workable answers.
- Read every explanation, including for questions you got right. A lucky guess teaches you nothing.
Ready to start?
Practice is free and needs no sign-up. Your progress is saved in your browser.