Advanced|19 hours|37 lessons

CI/CD Pipeline Engineering

Operating the build and delivery pipeline as a production system: feedback latency and where the minutes go, cache correctness, the test suite as infrastructure, autoscaling a runner fleet on Kubernetes, monorepo scale and merge queues, artifact promotion, and debugging a pipeline that fails only in CI. Lifetime updates included.

Text-based, no videos

9 modules, 37 lessons

Lifetime access

What you'll learn

Why feedback latency is the metric a pipeline is judged on, and where your minutes actually go

Cache keys as a correctness problem, and the failures a wrong key produces

Test sharding, test selection, and the arithmetic that makes flaky tests fatal at scale

Running and autoscaling an ephemeral runner fleet on Kubernetes against queue depth

Monorepo affected target selection, and merge queues that keep main green under load

Build once and promote, with an artifact identity that makes the claim provable

Debugging a failure that reproduces only in CI, and operating the pipeline as a service

Measuring delivery with the metrics as they are defined today, and how each one is gamed

Curriculum

9 modules · 37 lessons

What a Pipeline Actually Is

The pipeline as a feedback system with a latency budget, rather than a file of YAML.

4 lessons

The Pipeline Is a Feedback System22 minFREE Stages, DAGs and Gates26 minFREE Where the Minutes Actually Go30 minFREE Why Pipelines Rot24 minFREE

Making the Pipeline Fast

Parallelism, the critical path, and caching that is correct rather than merely fast.

5 lessons

The Cost of a Slow Pipeline28 min Parallelism and the Critical Path32 min What Makes a Cache Key Correct35 min When the Cache Is the Bug30 min Ordering Stages to Fail Fast26 min

Tests as Pipeline Infrastructure

Treating the test suite as a system you operate, with throughput, reliability and cost.

5 lessons

The Suite You Have and the Suite You Need28 min Sharding and Parallel Execution32 min Running Only What Changed35 min Flaky Tests and the Arithmetic of Retries38 min Quarantine and the Debt It Creates26 min

The Runner Fleet

Running the compute your pipeline executes on, including autoscaling it on Kubernetes.

5 lessons

Hosted, Self Hosted, and What You Are Buying26 min Runners on Kubernetes35 min Autoscaling Against Queue Depth32 min Warm Pools, Cold Starts and Image Pull30 min Runner Cost and Capacity Planning28 min

Monorepos and Merge Queues

What changes when one repository holds everything, and how main stays green under load.

4 lessons

When Run Everything Stops Working30 min The Dependency Graph and Affected Targets35 min Merge Queues and Why Green Pull Requests Break Main38 min Trunk Health as an Operational Metric27 min

Artifacts, Versions and Promotion

One artifact, many environments, and the identity that makes that claim provable.

4 lessons

Build Once, Promote Many28 min Versioning and Artifact Identity26 min Configuration Is Not the Artifact30 min Where the Pipeline Hands Off28 min

Pipeline Reliability and Debugging

The pipeline as a production system that fails, and how to debug one under pressure.

5 lessons

A Failure You Cannot Reproduce32 min Nondeterminism in the Pipeline35 min Retries as a Signal, Not a Fix27 min Debugging a Pipeline You Did Not Write30 min When the Pipeline Itself Is Down28 min

Pipelines at Organisation Scale

Hundreds of repositories, shared definitions, and measuring delivery without lying to yourself.

4 lessons

Reusable Definitions and Template Drift30 min Governance Without Blocking28 min Measuring Delivery Honestly35 min Migrating CI Systems Without a Freeze32 min

Capstone

A complete delivery platform design combining every module, with explicit tradeoffs.

1 lessons

Capstone: Design the Delivery Platform60 min

About the Author

Sharon Sahadevan

AI Infrastructure Engineer

Building production GPU clusters on Kubernetes. H100s, large-scale model serving, and end-to-end ML infrastructure across Azure and AWS.

10+ years designing cloud-native platforms with deep expertise in Kubernetes orchestration, GitOps (Argo CD), Terraform, and MLOps pipelines for LLM deployment.

Author of KubeNatives, a weekly newsletter read by 3,000+ DevOps and ML engineers for production insights on K8s internals, GPU scheduling, and model-serving patterns.

Ready to master this topic?

Start with the free preview lesson and see for yourself.