01
From one agent to a team
Orchestration patterns, when to split an agent and the coordination failure modes.
START FREE25 SLIDES · 50 MIN
AI Academy · AI-05 · Expert
Agentic Team on GCP: Run a whole team of agents in production.
AI-05 · EXPERTACADEMY ACCOUNT
Free
with a free accountNo cardOne account unlocks this course and every other one, in full.
- This course10 modules · ~14h
- Pass unlocksEvery course
- FormatSelf-paced + hands-on
- CardNone needed
Run a whole team of agents in production: Cloud Run, Firestore, scheduling, observability and security.
Who it is for: CTOs, senior developers and agencies running agents as infrastructure.
Syllabus
10 modules, ~14 hours, one capstone.
Module one is open to everyone, no account needed. The rest of the course opens the moment you create a free account, along with every other course.
02
GCP foundation
Projects, IAM, budgets and Secret Manager, set up the way production teams do it.
LOCKED23 SLIDES · 60 MIN
03
Compute for agents
Cloud Run services vs jobs vs worker pools, and which shape fits which agent.
LOCKED18 SLIDES · 60 MIN
04
State and queues
Firestore for agent memory, leases and queues that survive restarts.
LOCKED20 SLIDES · 60 MIN
05
Scheduling and events
Cloud Scheduler, Pub/Sub and webhooks: agents that wake up on their own.
LOCKED19 SLIDES · 55 MIN
06
Model strategy
Claude API and Vertex AI side by side: choosing per task, with fallback chains.
LOCKED20 SLIDES · 55 MIN
07
Observability and reliability
Logging, alerting, cost dashboards, retries and the outage playbook.
LOCKED21 SLIDES · 60 MIN
08
Security
Least privilege, sandboxing and prompt-injection defence for agents with real access.
LOCKED20 SLIDES · 60 MIN
09
Build weeks
A research agent, a content agent and a CRM agent sharing one pipeline.
LOCKED19 SLIDES · 150 MIN
10
Capstone
Your own three-agent team, live, plus an architecture review.
LOCKED19 SLIDES · 150 MIN
CAPSTONE
Your own three-agent team live on GCP, with an architecture review submission.
What you walk away with
Skills that you can use straight away.
An architecture for multi-agent teams: orchestration, state and queues
Agents deployed on Cloud Run services, jobs and worker pools
Firestore as agent memory with Cloud Scheduler and Pub/Sub driving work
Observability, budgets and model-fallback chains that survive outages
Security: least privilege, sandboxing and prompt-injection defence
TOOLS YOU WORK IN
Google Cloud RunFirestoreCloud Scheduler + Pub/SubSecret ManagerVertex AIClaude APIInside every lesson
Interactive learning activities.
Each lesson is a short deck of activity-book pages. You predict before you are told, do the task in your own store, and keep the notes. Built to be done, not skimmed.
Predict, then reveal
Commit a guess before the lesson gives you the answer, so it actually sticks.
Hands-on labs
Tick off real tasks in your own store, one at a time, with your progress saved.
Write-in workbook
Jot your answers straight into the page. They save on your device, yours to keep.
No walls of text
Flip-cards, compare panels and deep-dives instead of dense paragraphs to wade through.
Lab
LAB · AGENTIC TEAM ON GCP
- Do the first task in your real workspace
- Tick each step off as you complete it
- Save the result as your deliverable
YOUR CALL FIRST
What will you try first?
Before you ask
Questions, answered.
Running multiple AI agents as production infrastructure on Google Cloud: team architectures (pipeline, orchestrator, blackboard), IAM and Secret Manager foundations, Cloud Run services, jobs and worker pools, Firestore as agent memory with queues and leases, Cloud Scheduler and Pub/Sub, choosing between the Claude API and Vertex AI per task with fallback chains, observability and cost dashboards, and security including prompt-injection defence. The build weeks assemble a three-agent pipeline; the capstone puts your own team live.
CTOs, senior developers and agencies who want agents running as reliable infrastructure, not demos. You should be comfortable with cloud concepts and code; the Managed Agents course (or equivalent experience) is the assumed baseline.
Because the shape fits: Cloud Run gives each agent the right compute shape, Firestore holds shared state, Pub/Sub carries handoffs, Scheduler is the heartbeat, and IAM plus Secret Manager keep each agent's blast radius separate. It is also the stack we run our own agency agents on, so the course teaches from production scar tissue, including a real outage playbook.
Practice. D2C Prominence runs client audits, content, CRM and reporting on this exact architecture. The build weeks reproduce a working research-to-content-to-CRM pipeline modelled on ours, and the reliability module is built from outages we survived.
The GCP resources in the labs run comfortably inside the free tier plus a few dollars: Cloud Run scales to zero, Firestore's free quota covers the exercises, and the budgeting module sets spend alerts before anything else deploys.
Start now. Lesson one, the team-architecture deep dive, is readable with no signup. The rest opens with a free account, which covers every other course as well. No card, now or later.
AI-05 · FREE WITH AN ACCOUNT