AI Systems Engineering & Reliability

Format: Instructor-Led
Duration: 4 Days • 32 Hours

Overview

Learners will move from building and integrating AI solutions to confidently operating, monitoring, and scaling them in real-world environments. Through hands-on labs and live simulations, they’ll practise automation, reliability engineering, and troubleshooting to keep AI-enabled systems stable and efficient after deployment.

Outcomes

Deploy and manage AI-enabled systems in distributed, cloud-based environments (AWS, GCP, Azure).
Implement observability, monitoring, and alerting systems to track performance, drift, and uptime.
Apply reliability engineering principles, including redundancy, failover, and automated recovery.
Build and maintain CI/CD pipelines that support continuous integration and model updates.
Integrate testing and validation workflows for production-ready AI systems.
Diagnose and resolve incidents using root cause analysis and automated remediation tools
Optimise infrastructure performance, scalability, and cost-efficiency across compute and storage resources.
Collaborate across engineering, data, and AI teams to maintain resilient, compliant, and secure operations.

Prerequisites

None - There are no formal prerequisites.

This is an intermediate-level course. It is recommended that learners are familiar with the fundamentals of machine learning and Natural Language Processing as well as how AI applications are built and deployed. Familiarity with programming concepts, cloud environments, or prior experience (between 1-2 years) in data or software workflows will help learners get the most from the hands-on labs.

System Requirements

Below are details covering the course systems and accounts you need on Day 1. You’ll receive additional information about the rest of the required tools when the course begins. Your instructor may add or remove tools from your course. You will receive a full setup document as part of your course materials.

Technical requirements - operating system

The following OSs are supported in the Build AI Web Applications course:

Windows: The most recent public release of Windows 11.
macOS: The last two major public releases of macOS.
Ubuntu: The most recent LTS release of Ubuntu.

General Assembly can only offer support for the operating systems listed above. We do not offer support for beta or prerelease versions of any OS. We do not support any mobile/tablet operating systems or ChromeOS.

Students must have an account with administrator privileges on the computer they use for the course to install and configure the applications and tools required.

Technical requirements - hardware

All hardware should meet the minimum requirements to support the software requirements stated above, in addition to the following:

An Intel or AMD dual-core processor with at least a 2.0Ghz base clock speed or any Apple Silicon chip. Note that dual-core processors have a reduced frame rate when screen sharing on Zoom.
8GB of RAM.
Laptops should have at least a 13” screen size.
At least 40GB of free hard disk space.

Additionally, Windows and Ubuntu devices must be capable of enabling Intel Virtualization Technology (VT-x) or AMD Virtualization (AMD-V).

If taking the course in-person, a laptop equipped with Wi-Fi is required at every session.

Additional considerations for Build AI Web Applications

The following are all required at every session if taking the course remotely:

A functioning dual-monitor setup. At least one screen should be 27” or larger.
A functioning webcam (either built-in or external).
Access to stable, fixed broadband internet (100Mbps download / 20 Mbps upload).
A dedicated, quiet, well-lit workspace where you feel comfortable focusing. Preferably, this is a private room.

We highly recommend:

A functioning wired Ethernet connection.
Comfortable headphones.

Build AI Web Applications systems and tools

System	Needed Day 1?	Download / Access
Canvas	Yes	You will receive access after enrollment.
Google Chrome	Yes	Download Google Chrome.
Slack	Yes	Download the Slack app.
Zoom	Yes	Download the Zoom Desktop Client.
LLM	Yes	Choose one LLM: ChatGPT, Claude, or Gemini.
Git and GitHub	Yes	Create a free GitHub account.
JupyterLab	Yes	Install with Python and pip: `python -m pip install jupyterlab`, then launch with `python -m jupyter lab`.
Python 3.10	Yes	Download Python.
AWS Free Tier	Yes	Create a free AWS account.
Terraform	Yes	Download Terraform.
Docker Desktop	Yes	Download Docker Desktop.
Miro	Yes	Create a free Miro account.

Module Repos

#	Module Name	Module Overview	Associated Labs
1	Introduction to AI System Operations	Learn the fundamentals of maintaining and managing AI-enabled systems in production. Understand how data pipelines, models, and applications interact across the system lifecycle.	System Landscape Mapping Visualise how data, models, and services connect within an operational AI system. This lab helps learners understand dependencies and the flow between components in a live environment.
2	Cloud Environments and Infrastructure basics	Explore how cloud platforms support scalable AI applications and distributed workloads. Gain practical experience provisioning and managing resources using AWS free-tier services.	Build Your First IaC Template Create and apply a Terraform file to launch a small virtual environment on AWS. Learners practise automating infrastructure setup to ensure consistency and reproducibility.
3	Configuration Management and Version Control	Discover how infrastructure-as-code and version control tools ensure consistency and traceability in AI environments. Learn to automate setup, manage dependencies, and roll back safely.	Environment Setup Walkthrough Use Docker Compose to run multiple environments locally for development, testing, and production. This lab builds confidence in managing isolated configurations efficiently.
4	Principles of Reliability Engineering	Examine the principles of reliability engineering and how they apply to AI systems. Learn strategies to design for resilience, fault tolerance, and uptime in complex architectures.	Error Budget Calculator Model service reliability using SLIs, SLOs, and error budgets. Learners simulate trade-offs between innovation speed and reliability targets to guide engineering decisions.
5	Observability and Log Management	Develop observability skills through metrics, logs, and traces. Learn how to use monitoring tools to detect anomalies, visualise performance, and maintain transparency across AI pipelines.	Monitoring Dashboard Setup Collect and visualise live metrics with Prometheus and Grafana. This hands-on lab builds foundational skills in monitoring system performance and detecting anomalies.
6	Incident Response and Postmortems	Understand how teams prepare for, respond to, and learn from incidents. Practise documenting root causes and designing postmortems that drive system and process improvements.	Post-Mortem Simulation Review a mock incident and create an improvement plan. Learners practise identifying root causes, documenting lessons learned, and proposing reliability enhancements.
7	Continuous Integration Pipelines	Explore continuous integration pipelines and how they streamline model and system updates. Learn to automate testing, build verification, and deployment workflows for reliability.	Pipeline Builder Create a simple CI/CD pipeline that builds and deploys a Docker app automatically. Learners experience how automation reduces human error and improves release speed.
8	Container Deployment with Kubernetes	Deploy and manage AI containers using Kubernetes. Learn to orchestrate workloads, manage clusters, and optimise resources for scalability and performance.	Testing Workflow Setup Integrate automated tests into a deployment pipeline using PyTest and GitHub Actions. This lab ensures learners can verify AI model and API performance consistently.
9	Argo and Advanced Deployment Strategies	Implement deployment automation with ArgoCD and advanced rollout strategies. Gain hands-on experience managing versioned deployments, blue-green rollouts, and canary testing.	Canary Deployment Simulation Test a small rollout using ArgoCD and observe system responses. Learners gain confidence deploying updates safely and implementing rollback strategies.
10	Scaling Strategies for AI Systems	Investigate methods for scaling AI systems horizontally and vertically. Learn how to balance efficiency, cost, and performance across distributed environments.	Container Build Challenge Containerise a web app using Docker to ensure consistent execution across environments. Learners strengthen understanding of portability and reproducibility.
11	Load Balancing and High Availability	Examine load balancing techniques and strategies for ensuring high availability. Learn to design redundant, failover-ready systems for continuous AI service delivery.	Kubernetes Cluster Deployment Deploy and scale a service using Minikube and Kubernetes. This lab demonstrates orchestration fundamentals and how workloads are managed dynamically.
12	Security in AI Systems Operations	Understand the fundamentals of securing AI systems and protecting sensitive data. Learn to identify vulnerabilities, manage access, and apply DevSecOps principles.	Security Audit Simulation Learners identify and patch vulnerabilities in a sandboxed AI system environment. They apply DevSecOps principles to secure AI pipelines, protect data integrity, and implement monitoring for potential threats. The lab also introduces AI security wrappers and tools that help manage model access, prevent data loss, and defend against prompt injection and other AI-specific vulnerabilities.
13	Performance Testing and Tuning	Evaluate techniques for testing and tuning performance in AI applications. Learn to use benchmarking and profiling to improve responsiveness and reduce latency.	Cost Optimisation Dashboard Use Prometheus and Grafana metrics to tune system performance. Learners interpret real-time data to balance speed, cost, and reliability.
14	Cost Optimisation and Resource Management	Analyse how to optimise resource allocation and cloud costs without sacrificing performance. Learn to monitor usage patterns and apply automation for efficiency.	Workflow Automation Lab Build a lightweight, event-driven workflow using Apache Airflow. This lab reinforces automation principles for managing data pipelines efficiently.
15	Reliability Audits and Continuous Improvement	Learn to conduct reliability audits and develop continuous improvement cycles. Use metrics and retrospectives to track performance and drive operational excellence.	Failure Injection Exercise Simulate downtime with Chaos Mesh and document recovery results. Learners test resilience and learn how to strengthen fault tolerance in production environments.
16	Capstone Project: Operate, Secure and Test an AI System	Apply all learned concepts in a final capstone project. Operate, secure, and test a deployed AI system, demonstrating reliability, observability, and scalability in practice.	Capstone Project Plan, deploy, monitor, and test a full AI system end-to-end. Learners integrate reliability, automation, and observability to demonstrate operational excellence.