new series · hands-on projects
Cloud & AI Engineering
One AI application, grown from a single $5 server to a multi-cloud platform — deployed on AWS, Azure and DigitalOcean. Model serving, memory, traffic, GPUs, and the bill.
- AWS
- Azure
- DigitalOcean
- projects
- 9
- projects
- clouds
- 3
- clouds
- levels
- 3
- levels
- to follow along
- $0
- to follow along
# one app, nine rungs
You build one thing, then you make it survive real traffic
Every project grows the same application — Ask My Docs — a document question-answering API. It starts as one process on one server that a handful of people can use. It finishes as an autoscaled, observed, multi-cloud platform with GPUs that appear only when they are needed. Nothing is thrown away between projects, so you always know why the next piece exists.
- 01~10 people at once · $5–12 / month
- 02~50 people at once · $10–20 / month
- 03~50 people at once · $25–40 / month
- 04~500 people at once · $0 idle → ~$60 busy
- 05~5,000 people at once · $70–120 / month
- 06~50,000 people at once · $150–600 / month
- 07the same — but now you can see it · +$10 / month
- 08~50,000 people at once · measured against AWS
- 09~50,000, on any of three clouds · the full comparison
The projects, in order
Each one is a full build with a video and a repo. Open any project for the plan, the stack and what it costs to run.
- 01BeginnerRecording next
Your First AI API on One Cloud Server
One server, one model, one public URL — the thing you spend the rest of the series growing.
FastAPIUbuntuNginxsystemd~50 minOpen → - 02BeginnerComing soon
Containerize the Model — Docker & a Private Registry
Stop saying “it works on my machine” about four gigabytes of model weights.
DockerECRACRDO Registry~45 minOpen → - 03IntermediateComing soon
Give It Memory — Vector Storage & RAG
Where AI context and memory actually live, and what they cost you.
PostgrespgvectorEmbeddingsRDS~60 minOpen → - 04IntermediateComing soon
Managed Serving — Run It Without Managing a Server
Scale to zero when nobody is asking. The price is a cold start.
App RunnerContainer AppsDO App Platform~50 minOpen → - 05IntermediateComing soon
Traffic Management — Load Balancers, Streaming & Rate Limits
AI traffic is slow, bursty and expensive. Normal web defaults break it.
ALBAuto ScalingSSE streamingRedis~55 minOpen → - 06AdvancedComing soon
Autoscaling on Kubernetes — GPUs, Queues & Spot
AI workloads do not scale on CPU percent. They scale on queue depth.
EKSKEDAGPU nodesSpot~70 minOpen → - 07AdvancedComing soon
Observability & Cost Control for AI
The bill is a production metric. So is whether the answers are still any good.
PrometheusGrafanaTracingBudget alerts~55 minOpen → - 08AdvancedComing soon
Port the Whole Platform to Azure
Same container, same shape, different cloud. Here is what does not translate.
AKSFront DoorAzure PostgresTerraform~60 minOpen → - 09AdvancedComing soon
DigitalOcean & the Three-Cloud Head-to-Head
Third cloud, then the numbers: cost, cold start, GPU availability, and pain.
DOKSDO App PlatformDNS failoverTerraform~65 minOpen →

Muhammad Rashid
DevOps & Cloud Engineer · Instructor
I build and run production infrastructure, and I teach it the same way I do it — from a terminal, on real cloud accounts, with the bill visible. I founded CodeWithMuh and DevOps by Muh, where every course ends with something actually deployed rather than a certificate.
- free lessons published
- 30
- free lessons published
- end-to-end projects
- 13
- end-to-end projects
- practice questions written
- 299
- practice questions written
- teaching channels
- 2
- teaching channels
What you need first
No machine learning background. You do not train models in this series — you run them, and everything about running them is infrastructure. If you can use a terminal and Git, you can start at project 01.
Both are free courses on this site, with videos and notes.
Get each project as it lands
One email per project — the video, the repo, and the notes. No other mail.
$ subscribe --new-articles
Get new articles in your inbox
One email when a new hands-on guide goes live — Kubernetes, AWS, CI/CD, MLOps. No spam, unsubscribe anytime.