Home › Paths › Software Engineering
Learning path
Software Engineering
You build the services everything else runs on, and you make them dependable.
Why this path
Data, ML and AI work all stand on ordinary software: APIs, databases, queues, tests, pipelines and monitoring. Strong software engineering is the fastest way to become good at all the other paths.
How to use this page: 1. learn the key concepts, which are the same everywhere; 2. pick your platform; 3. work through Crawl, Walk and Run.
Read alongside: Agent Harness, drawn simply · Your first 30 days.
Step 1 · the same everywhere
Key concepts
Learn these ideas first. They are the same on every cloud and with every tool; only the product names change. Each one is a short page: what it is, why it matters, an example and the common mistake.
Working together
Building services
Running services
Step 2
Where do you work?
Pick your platform and you get one page with that platform's products in every step.
Beginner · 8 weeks
Crawl: You can build one thing end to end, with help.
Build and deploy a small service with tests and a pipeline.
Key ideas at this level, the same on every platform:
Open this level on: AWS · Azure · Google Cloud · Open source
Learn: the concept, and the product on each platform
| Learn | AWS | Azure | Google Cloud | Open source |
|---|---|---|---|---|
| A language, properlyfunctions, classes, errors, packaging | Python | Python | Python | Python |
| Git and pull requestsbranch, commit, review | GitHub | GitHub or Azure Repos | GitHub | GitHub |
| APIs and databasesREST, SQL, migrations | Lambda / ECS Fargate RDS Postgres / DynamoDB | Container Apps / Functions Azure Database for PostgreSQL / Cosmos DB | Cloud Run Cloud SQL / Firestore | Docker Postgres |
| Tests and CIunit tests, a pipeline that runs them | GitHub Actions + CodeDeploy | GitHub Actions or Azure Pipelines | GitHub Actions + Cloud Build | GitHub Actions |
Build on GitHub
Tasks API, done properly
A small REST API with a database, tests, a CI pipeline and a deploy. The boring basics, done right once.
You are ready to move on when
- Every change goes through a pull request with tests.
- You can deploy with one command, or none.
- Your README gets a stranger running in ten minutes.
Intermediate · 3 months
Walk: You can build it so it keeps working, without help.
Build services that survive failures: queues, retries, identity and monitoring.
Key ideas at this level, the same on every platform:
Open this level on: AWS · Azure · Google Cloud · Open source
Learn: the concept, and the product on each platform
| Learn | AWS | Azure | Google Cloud | Open source |
|---|---|---|---|---|
| Events and queuesdecouple services; retries; dead letters | SQS / EventBridge | Service Bus / Event Grid | Pub/Sub | RabbitMQ or Redis |
| Idempotencynever do the same thing twice by accident | RDS Postgres / DynamoDB | Azure Database for PostgreSQL / Cosmos DB | Cloud SQL / Firestore | Postgres |
| Identity and secretswho may call what; no keys in code | IAM + Secrets Manager | Entra ID + Key Vault | IAM + Secret Manager | Keycloak + sops |
| Observabilitylogs, metrics, traces; alerts that matter | CloudWatch + X-Ray | Azure Monitor + App Insights | Cloud Logging + Trace | OpenTelemetry + Grafana |
| Infrastructure as codethe environment in files | Terraform or AWS CDK | Bicep or Terraform | Terraform | Docker Compose |
Build on GitHub
Event-driven orders
Orders flow through a queue to payment and shipping. Retries, idempotency (never charge twice) and tracing across services.
Agent harness from scratch
Build the loop, tools, context and checks around a model in plain Python, one small step at a time.
You are ready to move on when
- A service can go down and no order is lost or doubled.
- You can find a slow request from one trace.
- No secret is in your repo.
Advanced · 6 months +
Run: You can design it, prove it and lead others through it.
Design systems and paved roads that other teams use.
Key ideas at this level, the same on every platform:
Open this level on: AWS · Azure · Google Cloud · Open source
Learn: the concept, and the product on each platform
| Learn | AWS | Azure | Google Cloud | Open source |
|---|---|---|---|---|
| System designtrade-offs, decision records, failure modes | Amazon EKS | AKS | GKE | kind or k3d |
| Platform engineeringtemplates and golden paths | Terraform or AWS CDK GitHub Actions + CodeDeploy | Bicep or Terraform GitHub Actions or Azure Pipelines | Terraform GitHub Actions + Cloud Build | Docker Compose GitHub Actions |
| Ports and adapterscode to interfaces; swap a vendor without a rewrite | Python | Python | Python | Python |
| Building with AIcoding agents, safely | Amazon Bedrock Bedrock tool use | Azure OpenAI in Microsoft Foundry Foundry function calling | Gemini on Vertex AI Gemini function calling | Ollama, or a free hosted open model plain Python loop, no framework |
Build on GitHub
Golden-path service template
A template repo any team clones to get a service with auth, logging, CI, infrastructure as code and a runbook on day one.
MyClaudeCode
A small coding agent built from scratch, to learn how context is budgeted and how tools are kept safe.
AI job-loss pain registry
Collect real problems people face from AI job loss and build small free tools that help.
Babu, a companion for elders
A tablet companion for an elderly person: reminders, conversation and a gentle eye for trouble.
You are ready to move on when
- Another team ships a service from your template.
- You can swap a cloud service by writing one adapter.
- You have led a design review.