Roadmaps › DevOps & Cloud
Cloud Engineer (AWS) Roadmap
Learn AWS from zero to the Solutions Architect Associate level — IAM, compute, storage, databases, networking, serverless and cost control.
Level: Beginner → Intermediate · Duration: 4–5 months
Step 1: Cloud fundamentals
What the cloud is and why companies use it.
Cloud computing means renting computing resources on demand instead of owning hardware. Understand the models and AWS's global layout before creating anything.
- Benefits: pay-as-you-go, elasticity, global reach
- Service models and serverless
- Regions, AZs and edge locations
- Safe account setup
Resources: AWS documentation · ▶ AWS Zero to Hero
Cloud models
IaaS, PaaS, SaaS, serverless.
The more AWS manages, the less you have to — but the less control you have.
- IaaS (EC2), PaaS (Elastic Beanstalk, RDS), SaaS
- Serverless (Lambda, DynamoDB)
- Public, private and hybrid cloud
Regions & availability zones
Where your resources live.
Where you put resources affects latency, cost, availability and legal compliance.
- Choosing a region (ap-south-1 for India)
- Spreading across AZs for high availability
- Edge locations for CloudFront
Shared responsibility
What AWS secures and what you secure.
AWS secures the infrastructure; you secure your data, identities and configuration.
- Security "of" the cloud vs "in" the cloud
- What changes for EC2 vs Lambda vs S3
Account setup & billing
MFA, budgets, the free tier.
Protect the account on day one — surprise bills and hacked root accounts are common beginner mistakes.
- MFA on the root user
- Budgets and billing alerts
- Free tier limits
- IAM Identity Center for daily logins
Step 2: Prerequisites
The basics every cloud engineer needs.
Cloud work builds on Linux and networking. Without them, EC2 and VPCs feel like magic.
- Linux administration basics
- IP addressing and subnets
- Scripting for automation
Resources: Linux documentation · ▶ Linux for Beginners
Linux
Terminal, SSH, permissions.
Most EC2 instances run Linux; you will SSH in and configure services.
- Shell commands and permissions
- Installing packages and running services
- Reading logs
Networking
IP addresses, subnets, DNS, ports.
VPCs, subnets, routes and security groups are networking concepts with AWS names.
- CIDR notation and subnetting
- Public vs private IPs, NAT
- DNS and ports
Resources: Notes: IP Addressing and Subnetting · Notes: Routing and the Network Layer
Scripting (optional)
Python or Bash for automation.
Automate repetitive tasks with the CLI and SDKs.
- AWS CLI with Bash
- boto3 with Python
Resources: Bash documentation
Step 3: IAM & security
Identity is the new perimeter.
IAM decides who can do what in your account. Most cloud breaches come from bad IAM, leaked keys or public buckets.
- Identities: users, groups, roles
- Writing least-privilege policies
- Protecting credentials
Users, groups & roles
Who can do what.
Give people groups; give services roles; avoid long-term keys.
- Groups for teams
- Roles for EC2, Lambda and cross-account access
- IAM Identity Center (SSO)
Policies
JSON permissions, least privilege.
Policies are JSON documents that allow or deny actions on resources.
- Effect, Action, Resource, Condition
- Managed vs inline policies
- Explicit deny wins
- IAM Policy Simulator / Access Analyzer
MFA & access keys
Protect accounts and credentials.
Credentials are the keys to the kingdom.
- MFA for every human user
- Never commit access keys
- Rotating and deactivating keys
Step 4: Compute
Run code on virtual machines, containers and functions.
AWS offers compute at every level of management — from raw VMs to functions that run for milliseconds.
- EC2 instances
- Scaling and load balancing
- Serverless with Lambda
- Containers on ECS/EKS
Resources: ▶ AWS Zero to Hero · ▶ Docker, Kubernetes & DevOps Courses
EC2
Instances, AMIs, key pairs, security groups.
EC2 gives you virtual servers with full control over the OS.
- Instance types and pricing models
- AMIs, key pairs and user data
- Security groups
- EBS volumes and snapshots
Auto Scaling & load balancers
ALB, NLB, scaling policies.
Scale out under load and replace unhealthy servers automatically.
- Launch templates and Auto Scaling Groups
- Application vs Network Load Balancer
- Target groups and health checks
- Scaling policies
Resources: Notes: Load Balancing
Lambda
Serverless functions and triggers.
Run code on events without managing servers; pay only when it runs.
- Handlers, runtimes and memory
- Triggers: API Gateway, S3, SQS, schedules
- Cold starts and limits
- Environment variables and layers
Containers
ECS, Fargate, ECR, EKS.
Run Docker containers without managing servers (Fargate) or with Kubernetes (EKS).
- ECR repositories
- ECS tasks and services on Fargate
- EKS basics
Resources: Docker documentation · Kubernetes documentation
Step 5: Storage
Files, disks and archives.
AWS storage comes in three shapes: objects (S3), blocks (EBS) and shared files (EFS).
- S3 buckets and objects
- Block and file storage
- Lifecycle and backup
S3
Buckets, storage classes, policies, pre-signed URLs.
S3 stores any amount of data with very high durability — uploads, backups, static sites.
- Buckets, objects and prefixes
- Storage classes
- Bucket policies and Block Public Access
- Pre-signed URLs and static website hosting
EBS & EFS
Block and shared file storage.
EBS is a disk for one instance; EFS is a network file system for many.
- EBS volume types (gp3, io2)
- Snapshots
- EFS mount targets
Backups & lifecycle
Versioning, Glacier, AWS Backup.
Automate cost-saving and protection.
- S3 versioning and lifecycle rules
- Glacier archive tiers
- AWS Backup plans
Step 6: Databases
Managed SQL and NoSQL.
Use managed databases so AWS handles backups, patching and failover.
- Relational: RDS and Aurora
- NoSQL: DynamoDB
- In-memory caching
Resources: Notes: Database Sharding and Replication
RDS & Aurora
Managed MySQL/PostgreSQL, Multi-AZ, read replicas.
Managed MySQL/PostgreSQL with automated backups and high availability.
- Creating an instance in private subnets
- Multi-AZ vs read replicas
- Parameter groups and snapshots
- Connecting from EC2 securely
DynamoDB
Serverless key-value at scale.
A serverless NoSQL database with single-digit millisecond latency at any scale.
- Partition and sort keys
- Designing for access patterns
- On-demand vs provisioned capacity
- Global secondary indexes
ElastiCache
Redis caching.
Managed Redis to cache queries and store sessions.
- Cache-aside pattern
- Cluster mode and replication
Resources: Redis documentation
Step 7: Networking in AWS
Your private network in the cloud.
A VPC is your private network in AWS. Good network design keeps databases private and apps reachable.
- VPC, subnets and routing
- Internet and NAT gateways
- DNS and CDN
- Connecting networks
VPC & subnets
Public and private subnets, route tables.
Plan address ranges and separate public and private tiers.
- CIDR blocks for VPC and subnets
- Route tables
- Security groups vs NACLs
Gateways
Internet gateway, NAT gateway.
Gateways control how traffic enters and leaves the VPC.
- Internet Gateway for public subnets
- NAT Gateway for private outbound access
- VPC endpoints for S3/DynamoDB
Route 53 & CloudFront
DNS and CDN.
Route 53 is DNS; CloudFront is the CDN that serves content from edge locations.
- Hosted zones and alias records
- Routing policies (latency, failover)
- CloudFront distributions with ACM certificates
VPN & peering (optional)
Connect networks together.
Connect VPCs to each other and to office networks.
- VPC peering
- Transit Gateway
- Site-to-site VPN
Step 8: Serverless & integration
Glue services together without servers.
Combine managed services to build apps with no servers to patch.
- HTTP APIs with API Gateway
- Queues, topics and event buses
- Workflows
Resources: Notes: Message Queues
API Gateway
HTTP APIs in front of Lambda.
Expose Lambda functions as REST or HTTP APIs with auth and throttling.
- HTTP API vs REST API
- Authorisers (JWT, Cognito)
- Stages and throttling
SQS, SNS & EventBridge
Queues, notifications, events.
Decouple services so a slow or failing part doesn't break the rest.
- SQS queues, visibility timeout, DLQs
- SNS fan-out
- EventBridge rules and schedules
Step Functions (optional)
Workflows.
Orchestrate multi-step processes with retries and error handling.
- State machines
- Retry and catch
Step 9: Monitoring, IaC & cost
Run it well and keep the bill low.
Production cloud work means observability, reproducible infrastructure and watching the bill.
- CloudWatch metrics, logs and alarms
- CloudTrail audit logs
- Infrastructure as code
- Cost optimisation
CloudWatch & CloudTrail
Metrics, logs, alarms, audit.
CloudWatch tells you how resources behave; CloudTrail records who did what.
- Metrics, dashboards and alarms
- Log groups and Logs Insights
- CloudTrail trails
Infrastructure as code
CloudFormation, CDK or Terraform.
Define infrastructure in code to recreate environments reliably.
- CloudFormation templates and stacks
- AWS CDK in TypeScript/Python
- Terraform as a multi-cloud option
Cost optimisation
Right-sizing, Savings Plans, cleanup.
Cloud costs grow quietly. Review them monthly.
- Cost Explorer and budgets
- Right-sizing and stopping idle resources
- Savings Plans, Reserved and Spot instances
- Tagging for cost allocation
Step 10: Certification & projects
Prove your skills.
Certifications help get interviews; projects help you pass them.
- Cloud Practitioner
- Solutions Architect Associate
- Portfolio projects
Resources: AWS Certification
Cloud Practitioner
The entry-level exam.
A broad, beginner-friendly exam about AWS services, billing and security.
- Core services overview
- Pricing and support plans
- Practice exams
Solutions Architect Associate
The most valued associate exam.
The most requested AWS certification — designing resilient, secure, cost-effective systems.
- Designing multi-tier architectures
- Choosing storage and databases
- High availability and disaster recovery
- Scenario-based practice questions
Projects
Static site on S3 + CloudFront, a 3-tier app, a serverless API.
Build real things and document the architecture with a diagram.
- Static website on S3 + CloudFront + Route 53
- Three-tier app: ALB + EC2/ECS + RDS
- Serverless API: API Gateway + Lambda + DynamoDB
- Everything defined with IaC