Viktor Farcic vfarcic

Intro

AI agents are amplifiers. If you're good at your job, agents make you better. You do more great things, faster. But if you're bad at your job, agents amplify that too. Where you used to cause a slow trickle of shit, now you have the means to unleash a full-blown shitstorm, at scale, in minutes.

Now, AI is all the rage these days, and for good reason. So of course people are using agents to manage real resources: infrastructure, databases, applications, all of it. The question is what happens when they do. That's what we're looking at today: an agent managing actual cloud resources, what goes wrong, why it goes wrong, and what it takes to make it work properly.

Setup

git clone https://github.com/vfarcic/infra-with-ai

name	manuscript-broll-suggest
description	Scan a manuscript for overlay-free narration segments and suggest cinematic b-roll concepts to generate via Higgsfield AI; on selection, generate the clip into ./tmp and open it in QuickTime, then offer to refine or pick another.

Suggest and Generate B-Roll for Manuscript via Higgsfield

Analyze a manuscript, find segments that have no other visual overlay, propose movie-like b-roll concepts for them, and on user selection generate the clip with Higgsfield AI, save it under ./tmp/, and open it in QuickTime. Then loop: refine the current clip or pick another suggestion.

Input

name	manuscript-broll-opener
description	Generate a cinematic 'cold open' clip that ends precisely at the first frame of recorded talking-head footage. Two-phase: brainstorm effect + suggest intro-narration tweaks to tighten the seam + give recording instructions, then resume with the user's first-frame screenshot to generate via Higgsfield with end_image. Tracks accepted concepts across videos to avoid repeats.

Generate Cold-Open Clip for Manuscript via Higgsfield

Propose a cinematic intro effect (materialize, ceiling zoom, rack focus, lamp reveal, etc.) for the very start of a video. The clip plays before recorded narration and ends at the exact first frame of the recorded talking-head footage, creating a seamless cut into live narration.

The skill runs in two phases:

Intro

TODO: Logo: https://openai.com TODO: Logo: https://anthropic.com TODO: Logo: https://deepmind.google

Every prompt you send to ChatGPT, Claude, or Gemini runs on someone else's GPUs. Your data, your code, your company's secrets, all flowing through infrastructure you don't control. For most people, that's fine. But if you're in healthcare, finance, government, or anywhere that compliance actually matters, "fine" isn't good enough.

Here's the thing. Running AI models on your own infrastructure sounds like it should be straightforward. It's not. GPUs are brutally expensive. Get the setup wrong and you're burning hundreds of thousands of dollars on mistakes. Get it right, and you have Inference-as-a-Service that any team in your company can use, with your data never leaving your network.

Intro

I've spent the last few months building custom AI agents for internal teams, and I can tell you: the gap between using a general-purpose coding assistant and having an agent that truly understands your company is enormous. We're talking about the difference between an intern who's brilliant but knows nothing about your business, and a senior engineer who's been with you for years.

In this video, I'll walk you through the complete architecture for building your own AI agent. Every component: system context, tools, knowledge retrieval, multi-agent orchestration, security, observability, and cost optimization. By the end, you'll have a blueprint you can actually follow. Not just theory, but the real decisions you'll face and how to make them.

Why Build Your Own Agent?

Let's say you're working in a company with a bunch of software engineers, and everyone wants to use AI to help with coding, operations, debugging, you name it. The obvious move is to just hand everyone Claude Code, Cursor, or Windsu

You are analyzing documentation to identify all content that can be validated through testing. Your goal is to find every section containing factual claims, executable instructions, or verifiable information.

File to Analyze

File: {filePath} Session: {sessionId}

Core Testing Philosophy

Most technical documentation is testable through two validation approaches: :...skipping...

Vector Databases, Embeddings, and RAG: A Practical Guide

Introduction

What are Vector Databases, Embeddings, and RAG?
Why they matter in modern AI applications
How the DevOps AI Toolkit uses these technologies for intelligent pattern matching

Core Concepts

Manual (Without GitOps)

go test --tag unit

docker image build --tag ghcr.io/vfarcic/silly-demo:v1.2.3 --push

yq --inplace ".spec.template.spec.containers[0].image = \"ghcr.io/vfarcic/silly-demo:v1.2.3\"" staging/deployment.yaml

kubectl --namespace staging apply --filename dev/deployment.yaml

Introduction

Imagine if you could create, for people in your company, a platform that would provide them with the same experience they have when working with AWS, Google Cloud, Azure, or any other public Cloud provider. Imagine if there would be a service for everything they do.

Do you need a database that works exactly as we expect it to work in this company with all the security, backup, compliance, and other policies we have?

Well...

There is a service for that.

	## Setup

	```sh
	git clone https://github.com/vfarcic/infra-with-ai

	cd infra-with-ai
	```

	> Make sure that Docker is up-and-running. We'll use it to create a KinD cluster.