Training on building, securing and running software with AI coding agents.

Self-paced lessons from Intellimetrics on directing AI agents from Claude Code, OpenAI Codex, Google Antigravity and GitHub Copilot, whether you start in a desktop app, a terminal CLI or a cloud builder. Every path leads to the same skill, and the exercises check your work along the way.

Free · No sign-up · 11 trainings · 2 pathways · 272 lessons · about 78 hours

Enter the code from your program coordinator to open your organization's track.

The fundamentals

AI-Powered Development teaches the AI coding tools named above and how to plan, direct and check an agent's work. The trainings below assume you can use one of them, except Leading in the AI Era, which needs no tool. Pick one of the 3 paths, based on your background. Each path begins with a first session of 5 lessons, about 85 to 95 minutes: the lessons to take first, each of them also part of the path's full track.

Self-paced · 13 modules · 58 lessons · about 16 hours

The trainings

9 numbered trainings use one fictional utility, Meridian, as a shared case study. Each stands on its own. The number on a card is the suggested order if you take more than one. Every one of them has a one-page leadership brief.

Part 1: the secure baseline

6 live

Cloud migration, Zero Trust, delivery pipelines, federal assessment, data products and document search.

☁️
01

Assess the legacy estate, choose a platform with managed services first, containerize, migrate the data, and get workloads ready for production.

Self-paced · 6 modules · 22 lessons · about 7 hours

Application teams, architects, SRE/operations

🔒
02

Zero Trust built from the federal guidance: identity-based access, device and network policy, data controls, logging that holds up as evidence, and the rollout.

Self-paced · 6 modules · 21 lessons · about 6 hours

Optional preparation: Cloud Modernization Patterns

Security teams, network engineers, architects

🔄
03

Deliver as one team, with security in the pipeline from the start. Trunk-based development, pipeline templates with the security gates built in, infrastructure as code, release practice, and build provenance an assessor can check.

Self-paced · 6 modules · 22 lessons · about 6 hours

Optional preparation: Cloud Modernization Patterns · Zero Trust Implementation

Software engineers, platform teams, security engineers, tech leads

🛡️
04

Turn compliance requirements into engineering work: control patterns, evidence collection, secure pipeline design, and documentation for assessment under the NIST Risk Management Framework.

Self-paced · 6 modules · 20 lessons · about 6 hours

Optional preparation: Zero Trust Implementation · Modern DevSecOps Foundations

Federal program teams, ISSOs, DevSecOps teams

📊
05

Turn data you were handed into something people can rely on. Define the question before you query, profile the data, make it rerun the same way, check it before you publish, and own it afterwards.

Self-paced · 6 modules · 22 lessons · about 7 hours

Optional preparation: Cloud Modernization Patterns · Modern DevSecOps Foundations

Analysts, domain experts, data stewards, and the engineers who support them

📄
06

Build a question-answering system over a governed set of documents, and learn what its answers do and do not prove.

Self-paced · 6 modules · 22 lessons · about 7 hours

Optional preparation: Building Trustworthy Data Products · Zero Trust Implementation

Analysts, domain experts, records owners, and the engineers who support them

Part 2: govern, operate, automate

3 live

AI release decisions, incident response and reporting, and agents that act within limits.

🧪
07

List every AI system, tier it by consequence, assemble the evidence, evaluate it on your own data, and put the release decision on a signed record that expires.

Self-paced · 6 modules · 23 lessons · about 7 hours

Optional preparation: Grounded Answers From Documents · Zero Trust Implementation

AI governance owners, ISSOs, security teams, engineers shipping AI features

📟
08

On-call, incident command, reporting clocks that each start at their own recorded trigger, communications, and post-incident reviews that lead to changes.

Self-paced · 6 modules · 23 lessons · about 7 hours

Optional preparation: Cloud Modernization Patterns · Modern DevSecOps Foundations · AI Assurance: System Risk and Release Decisions

SRE and operations, on-call engineers, service owners, incident-response leads

🤖
09

How to let agents take operational actions: each action approved on its own, inside limits the agent cannot change, and recorded where the agent cannot write.

Self-paced · 6 modules · 22 lessons · about 6 hours

Optional preparation: Operating in Production: On-Call, Incident Command, and Reporting Clocks · AI Assurance: System Risk and Release Decisions · Zero Trust Implementation

Platform and SRE teams, security engineers, operations leaders

Standalone trainings

3 live

Unnumbered courses and pathways. A pathway is a set of lessons from other trainings, in a set order, with a skip test at each stage, and adds no new lessons.

🧭

Leading in the AI Era

Live

Decide where AI is worth trying, what evidence and baseline to require before funding a build, how to clear the bottlenecks that stall useful work, and how to support the people doing it. No technical background or coding tool required.

Self-paced · 6 modules · 17 lessons · about 3 hours

Executives, directors, program managers, mission owners and academic leaders

🎛️

The two advanced modules of AI-Powered Development: context engineering, tool servers, skills and hooks, subagents, headless CI, spec-driven work, and review by several models. Each stage has a skip test.

Self-paced · 5 stages · 17 existing lessons · about 5 hours

Teams that ship code and want agents in the pipeline; tech leads approving the tooling

🏗️

Decide whether the demo should exist, prove its answers, name its consequence tier, evaluate it on your own data, run it with a pager and a record, and only then let it act. Whole modules from the Meridian trainings, with a skip test at each stage. It teaches no Python and no API integration.

Self-paced · 6 stages · 32 existing lessons · about 10 hours

Engineers who have built an AI demo and now have to run it for other people

Organization tracks

Each organization track pairs the fundamentals with a module for one field. All 8 open with a code from a program coordinator. Government Dev publishes its code, so anyone can try it.

Government Dev

Live Open to everyone

Development for a government agency with AI coding agents. Deploy a Next.js reference app to Azure Government, AWS GovCloud or Google Cloud, then produce SBOMs, write control statements, modernize legacy code, and add AI features a person can overrule.

Self-paced · 39 lessons after the fundamentals · about 29 hours

Developers building for a government agency · runnable reference app on GitHub (MIT license)

Open the track Code GOVDEV2026

7 more organization tracks

Built for university, agency and company cohorts. Each opens with the code your program coordinator gave your group.

For teams

For leaders deciding whether to send a team: every numbered Meridian training has a one-page leadership brief, linked on its card. Highlight the briefs

If you lead the work rather than do it, Leading in the AI Erais the course written for you. To set up a cohort or get access codes for your agency, campus or company, email contact@intellimetrics.net with your group's name, its field and a start date. Your group's track is the fundamentals plus a module for that field.