This MLOps training teaches you how machine learning models are trained reproducibly, versioned and monitored in production, and it is aimed at data scientists, ML engineers and DevOps teams. You work with experiment tracking, a model registry, automated pipelines and monitoring for data drift. Python knowledge and a first trained model are expected. We plan scope and tooling to fit your platform when you enquire.
What you’ll take away.
- Record experiments, parameters and metrics traceably with a tracking tool such as MLflow
- Version models in a registry and define promotion stages
- Build training and delivery steps as an automated pipeline with CI/CD
- Monitor data drift and model quality in production with metrics and alerts
- Protect new model versions with shadow deployment, canary rollout and rollback
Your training in detail
- Who is it for?
- Data scientists, ML engineers and DevOps and platform teams who want to bring models into production reliably.
- Prerequisites
- Python skills, basic Git and container knowledge, and an understanding of training and evaluating a simple model.
- Format & planning
- Suggested format: 3 days live online or on site, split into modules. Scope depends on prior knowledge.
Agenda: from understanding to application
Our independently developed agenda proposal. Focus areas, lab environment and scope are agreed before booking.
Lifecycle of an ML system and automation maturity levels following Google Cloud's MLOps guidance.
Reproducibility with Git, data versioning (for example DVC), pinned environments and seeds.
Experiment tracking and model registry with MLflow: runs, artefacts, model versions and aliases.
Pipelines and CI/CD for training, data and model tests, and automated delivery as a container.
Serving as a batch or online service, latency, scaling and rollout strategies such as canary and shadow.
Monitoring of input data, predictions and business metrics, drift detection and retraining.
Suggested exercise: You train a scikit-learn model, log runs in MLflow, register the best version and serve it behind a small API whose input drift you simulate.
Exams, licences & credentials
There is no single vendor-neutral MLOps exam. Cloud providers offer their own certifications with MLOps content, which you can check with them. This training does not prepare for them specifically, and attending alone awards no vendor certificate. Exam fees are only included if the offer confirms them.
Vendor references & review date
Sources checked on 2026-11-03. These sources describe the technology or external exam requirements. They do not imply a partnership with NextNowa.
Your learning content
These objectives describe the training topic. The specific syllabus, duration, dates, trainer and price are agreed before booking.
Frequently asked questions
What is the difference between MLOps and DevOps?
DevOps automates building and running software. MLOps extends this to data and models: training data, experiments and model quality must also be versioned and monitored, because a model can degrade without any code change when the data shifts. That is why drift monitoring and retraining come in.
Do I need Kubernetes for MLOps?
Not necessarily. For small applications, containers on a single server or a managed service are often enough. Kubernetes pays off when you must scale and isolate many models. The training concentrates on versioning, pipelines and monitoring, which are needed on any platform.
How do I recognise data drift in a model?
You compare the distribution of current input data with the training data, for example using statistical tests or per-feature metrics. You also watch the prediction distribution and, once available, the actual hit rate. If something deviates clearly, you investigate causes and possibly trigger retraining.
Does MLOps: operating models include a certificate?
There is no single vendor-neutral MLOps exam. Cloud providers offer their own certifications with MLOps content, which you can check with them. This training does not prepare for them specifically, and attending alone awards no vendor certificate. Exam fees are only included if the offer confirms them.