Save on courses or Exam now
About Instructors
-
edwardLary- Instructor Rating - Reviews - Students - CoursesNo biography available, please complete
🚀 Course Goals
- Deploy and operate a self-hosted n8n + PostgreSQL stack with Docker Compose, managed secrets, backups and a verified restore. Build workflows that hold up in production: expressions, branching, sub-workflows, batching, schema validation, idempotency keys and safe retries with backoff. Integrate real REST APIs end to end — webhooks, OAuth and token auth, pagination, rate limits, binary files and duplicate-event protection. Drive LLMs from a workflow and get reliable results: prompt design for machine consumers, structured JSON output, validation, extraction, classification and summarisation.
- Build tool-using AI agents with memory, bounded tool access, error handling and human-in-the-loop approval before any irreversible action. Ship a RAG knowledge assistant: ingestion, chunking, embeddings, a vector store, retrieval quality and grounded answers that cite their sources. Orchestrate multi-agent work with routers, orchestrator-worker sub-workflows, state machines, checkpoints, escalation and a full audit trail. Expose and consume tools over MCP with explicit permission boundaries and prompt-injection defences.
- Automate real business processes: lead intake and routing, CRM pipeline updates, support triage with SLAs, marketing approval flows, invoices, expenses, procurement and HR onboarding. Run n8n in production: domain and TLS, reverse proxy, queue mode with Redis workers, scaling, credential rotation, upgrades, rollback and disaster recovery. Operate what you built: correlation IDs, central error workflows, alerting, regression tests, LLM and agent evaluation, guardrails, latency, token use and cost control. Deliver a complete capstone — the Northstar AI Automation Operations Hub — with data contracts, acceptance tests, a business demo and a portfolio package.
📘 About This Course
Most automations do not fail on the day they are built. They fail three weeks
later, when the same webhook is delivered twice and a customer is refunded twice, when a partner
API starts returning 429 and the workflow retries in a tight loop, when an LLM
returns prose instead of the JSON the next node expects, or when an agent decides on its own to
send an email no human ever approved.
This course builds the automation platform that survives those days. You start from an empty
Ubuntu machine and finish with a running, auditable system: a self-hosted n8n and PostgreSQL stack
in Docker Compose, a local business sandbox with mail, a mock CRM, files and webhooks, workflows
that transform and validate real data, resilient API integrations with pagination, backoff and
idempotency keys, LLM steps that return validated structured output, tool-using AI agents with
memory and human-in-the-loop approval, a RAG knowledge assistant that answers only with sources,
multi-agent orchestration with audit trails, MCP interoperability with explicit permission
boundaries, and a production deployment with TLS, queue mode, Redis workers, backups, monitoring
and cost control.
Every lesson is a real session on a real Linux lab, not slides about one. You see the actual
editor, the actual execution list, the actual JSON moving between nodes, the actual failure and
the actual repair. The recurring company, Northstar Commerce, keeps leads,
customers, tickets, orders, invoices and policy documents coherent from the first lesson to the
capstone, so each module extends one system instead of starting over.
The whole stack is free and open source, runs on a single machine without a GPU and without a
paid cloud account, and every dataset is fictitious or sanitised. Optional SaaS integrations are
shown only where you can supply your own authorised credentials.
15 modules, 99 lessons, just under 20 hours, ending with the Northstar AI Automation
Operations Hub: an end-to-end capstone you deploy, harden, monitor and demonstrate.
What You'll Learn
Deploy and operate a self-hosted n8n + PostgreSQL stack with Docker Compose, managed secrets, backups and a verified restore.
Build workflows that hold up in production: expressions, branching, sub-workflows, batching, schema validation, idempotency keys and safe retries with backoff.
Integrate real REST APIs end to end — webhooks, OAuth and token auth, pagination, rate limits, binary files and duplicate-event protection.
Drive LLMs from a workflow and get reliable results: prompt design for machine consumers, structured JSON output, validation, extraction, classification and summarisation.
Build tool-using AI agents with memory, bounded tool access, error handling and human-in-the-loop approval before any irreversible action.
Ship a RAG knowledge assistant: ingestion, chunking, embeddings, a vector store, retrieval quality and grounded answers that cite their sources.
Orchestrate multi-agent work with routers, orchestrator-worker sub-workflows, state machines, checkpoints, escalation and a full audit trail.
Expose and consume tools over MCP with explicit permission boundaries and prompt-injection defences.
Automate real business processes: lead intake and routing, CRM pipeline updates, support triage with SLAs, marketing approval flows, invoices, expenses, procurement and HR onboarding.
Run n8n in production: domain and TLS, reverse proxy, queue mode with Redis workers, scaling, credential rotation, upgrades, rollback and disaster recovery.
Operate what you built: correlation IDs, central error workflows, alerting, regression tests, LLM and agent evaluation, guardrails, latency, token use and cost control.
Deliver a complete capstone — the Northstar AI Automation Operations Hub — with data contracts, acceptance tests, a business demo and a portfolio package.
👤 Who This Course is For
- Automation engineers and developers who can wire two apps together but want workflows that survive duplicates, rate limits, schema drift and incidents.
- Operations, revenue and back-office professionals who want to automate real processes with AI without handing irreversible actions to a model.
- Engineers evaluating AI agents for production who need approval gates, guardrails, evaluation and cost control rather than demos.
- Consultants and freelancers who want a self-hosted, auditable automation stack they can deploy for clients without a paid SaaS dependency.
53618 Students applied
100 Comments
-
-
John DoeMorbi nec nisi ante. Quisque lacus ligula, iaculis in elit et, interdum semper quam. Fusce in interdum tortor. Ut sollicitudin lectus dolor eget imperdiet libero pulvinar sit amet
-
John Doe
Morbi nec nisi ante. Quisque lacus ligula, iaculis in elit et, interdum semper quam. Fusce in interdum tortor. Ut sollicitudin lectus dolor eget imperdiet libero pulvinar sit amet
-
-
John Doe
Morbi nec nisi ante. Quisque lacus ligula, iaculis in elit et, interdum semper quam. Fusce in interdum tortor. Ut sollicitudin lectus dolor eget imperdiet libero pulvinar sit amet
-
John Doe
Morbi nec nisi ante. Quisque lacus ligula, iaculis in elit et, interdum semper quam. Fusce in interdum tortor. Ut sollicitudin lectus dolor eget imperdiet libero pulvinar sit amet
-
John Doe
Morbi nec nisi ante. Quisque lacus ligula, iaculis in elit et, interdum semper quam. Fusce in interdum tortor. Ut sollicitudin lectus dolor eget imperdiet libero pulvinar sit amet
Morbi nec nisi ante. Quisque lacus ligula, iaculis in elit et, interdum semper quam. Fusce in interdum tortor. Ut sollicitudin lectus dolor eget imperdiet libero pulvinar sit amet