Agentic Engineering

Vibe coding gets you to a working application in hours by describing what you want. This course takes you deeper: the mechanics underneath every agent, the patterns that make them reliable at scale, and the engineering discipline that turns a prototype into a production system you would trust with real users and real data.

If the Vibe Coding course is about building fast with AI, this course is about understanding what is happening under the hood so you can build things that do not break. The agent loop. API integration with retries and backoff. Memory systems built on vector databases. Prompt injection defense at the protocol level. Multi-agent coordination. Spec-driven contracts. Production deployment with observability.

You will write code. You will design systems. You will understand why the agent did what it did, not just that it worked.

Prerequisites

This course assumes you have completed Vibe Coding or have equivalent experience — you are comfortable with the terminal, you have built things with Claude Code or similar tools, and you are ready to go deeper. If you are new to the terminal or Python, start with Practical Fundamentals and Vibe Coding first.

Course structure

Day 1 -- The Agent Loop starts with the mechanical difference between a chatbot and an agent: the loop. You will build a working agent from twenty lines of Python, set up your development toolkit, and learn the prompting and context-window fundamentals that make everything downstream work.

Day 2 -- Tools & Interoperability is where an agent stops being a single model call and starts being a system. API integration with retries and backoff. Safe code execution in sandboxed environments. Event-driven automation. Agent-to-agent communication through MCP.

Day 3 -- Memory, Skills & Multi-Agent Systems gives your agent persistent memory across conversations using pgvector, teaches it to write and reuse its own skills, and shows you how to coordinate multiple agents in a single workflow.

Day 4 -- Security, Reliability & Evaluation is the day that determines whether what you built is a toy or production-ready. Structured output with schema validation. Prompt injection defense. Real evaluation with metrics instead of eyeballing outputs.

Day 5 -- Production Systems takes everything built over the previous four days and ships it: contract-first design, deployment that survives crashes, user-facing APIs, streaming interfaces, and observability with tracing and dashboards.

Final build

Every day feeds directly into the Final Build: The Foreman -- an on-call engineering assistant you page from Slack, built from two cooperating agents that research, fix, and remember. It is not a toy. It is the proof that everything you learned composes into a working production system.