Building ZeroCommit From LLM Response to Codebase in Seconds.
Joined January 2026
- Tweets350
- Following690
- Followers63
- Likes2.1K
Saad Khan retweeted
AWS Core Architecture - Simplified
Account
↓
Region
↓
VPC (network boundary)
↓
Subnets (public / private)
↓
Route Tables (traffic rules)
↓
Internet Gateway (internet access)
↓
NAT Gateway (outbound only)
↓
Security Groups (access control)
↓
Load Balancer (ALB / NLB)
↓
Target Groups
↓
Auto Scaling Group
↓
Compute (EC2 / Containers / Lambda)
↓
Application Code
↓
Datastores (RDS / DynamoDB / S3)
↓
IAM (access control)
↓
Logs (CloudWatch)
↓
Metrics (CloudWatch)
↓
Alarms
↓
Multi-AZ
↓
Backups
↓
Infrastructure as Code
↓
Cost Controls
↓
Operations
↓
Failures & recovery
🎯 If you understand this flow, AWS stops feeling random. It becomes predictable.
#AWS
Saad Khan retweeted
You are not a Cloud Architect if you are just:
❌ Deploying apps to AWS/Azure
❌ Using managed services like RDS or Lambda
❌ Following tutorials to spin up infrastructure
❌ Clicking through cloud consoles
❌ Copy-pasting Terraform from Stack Overflow
You are a Cloud Architect if you are:
✅ Designing multi-region failover strategies
✅ Building cost-optimized infrastructure at scale
✅ Creating security frameworks and compliance strategies
✅ Planning disaster recovery with RTO/RPO targets
✅ Capacity planning for unpredictable growth
✅ Making trade-off decisions between services (and defending them)
✅ Designing network architecture (VPCs, subnets, peering, gateways)
✅ Writing enterprise-level IaC with modules and state management
Cloud engineers deploy to the cloud.
Cloud Architect: Designs systems for scale, cost, security, and resilience
Saad Khan retweeted
AWS EC2 - In Plain English ☁️
Instances = Virtual Servers
→ On-Demand | Spot | Reserved
Instance types:
→ t → cheap, burstable
→ m → general purpose
→ c → compute heavy
→ r → memory heavy
AMIs:
→ Templates to launch instances
Auto Scaling Groups:
→ Automatically add/remove instances
→ Based on demand
EBS:
→ Persistent block storage
→ gp3 / io2
Mistakes that hurt 💸
→ No termination protection → accidental deletes
→ Oversized instances → wasted money
→ Lost SSH key → locked out forever
→ Idle instances running for months
→ No IMDSv2 → metadata exploit risk
EC2 = Power + Responsibility
→ Tag everything.
→ Monitor always.
First painful EC2 lesson you learned? 👇
