Explore software development, AI applications, and technical innovation at Sergei Liebich Tech Hub. As an experienced software developer, I share insights into modern web and mobile development, artificial intelligence, and cutting-edge technology solutions.
import 'package:flutter/material.dart';
import 'package:flutter/App/app.dart';
import 'package:hooks_riverpod/hooks_riverpod.dart';
void main() {
WidgetsFlutterBinding.ensureInitialized();
WidgetsFlutterBinding.ensureInitialized();
runApp(
const ProviderScope(
child: MyApp()
),
);
}
I'm a fullstack developer (Flutter, Nuxt, Spring Boot) writing about agentic coding in practice: how I build software with AI agent teams in Claude Code and GitHub Copilot — what works, what breaks, and what it costs.
None of it is theory. The AI agents work on my real projects — Flutter apps like Atomin, the habit tracker I built and ship, and Nuxt web apps like this blog. Every post comes from code that actually goes to production.
Alongside the agentic coding series you'll find my products, project write-ups, and tutorials on web and mobile development. Start with the blog, or take a look at what the agents and I have built so far.
11/15/2026
We review code and approve requirements with a handshake. Why I put a dedicated reviewer agent in front of every production stage, why an agent must never check its own output – and what this pattern costs me per feature.
11/08/2026
A structure nobody checks is just a request. How I turned my adapter files into a machine-verified contract, why the gate numbers stay identical across all projects – and why every weakening needs a target date.
11/01/2026
Loading domain knowledge only on demand sounded like the next big lever. An experiment with a stop threshold set in advance said no. What helped instead, and what the startup load in two real projects shows today.
10/25/2026
Feels expensive
10/18/2026
I develop seven projects on four stacks on my own, all with the same agent workflow. How that came about, why I had to split the workflow into core and adapter to make it work, and what six weeks of session data show about my share of the work.
09/20/2026
Four assistants, four different tasks — and the same sentence at the end of each: only as good as what is connected. Why better prompting rescues nothing here, how I connected missing knowledge first with local files and later via MCP — and why an agent with many sources needs a signpost.
09/13/2026
When a business test scenario turns red in the pipeline and a dozen services deploy at the same time, finding who caused it is harder than fixing it. Why an assistant's diagnosis stands and falls with logs and traces – and why it still has to deliver three hypotheses instead of one answer.
09/06/2026
Why reviewing is more expensive than creating, why a review assistant has to be able to say "I couldn't check that" – and why duplicate detection fundamentally doesn't work without full access to the existing inventory.
08/30/2026
Why the ever-identical conversion of one artifact type into the next is the most rewarding case for an assistant, why enriching and transforming have to be two separate steps – and why every enriched statement has to carry its origin.
08/23/2026
How assistants turn a vague requirement into a structured artifact, why acceptance criteria are the hardest part, and why an assistant has to surface assumptions instead of making them.
08/10/2026
The psychological flip side of the AI hype – FoMO-AI, the fear of being left behind, oversight labor and knowledge hiding. An honest look at why the feeling is real, and why you still aren't lagging behind.
08/03/2026
After the switch to AI credits, I hit my limit fast. How a local setup of Qwen3.6-27B, OpenCode, the Language Server Protocol and CodeGraph got me back up to speed – and saves 69 % of tokens according to CodeGraph's own benchmark.
07/27/2026
Agent files grow with every project. How I move recurring specialized knowledge into Agent Skills, what progressive disclosure actually saves – and why more skills don't automatically mean better agents.
07/13/2026
Which risks can arise from using AI agents in the development process? An honest look at skill atrophy, cognitive debt, hidden costs – and why the problem is real, but perhaps temporary.
07/06/2026
Which metrics can be used to measure the positive effect of AI agents in the development process? My approaches to agent performance measurement.
06/29/2026
How I integrate Figma and Appwrite MCPs into my agent workflow to cleanly connect frontend and backend without manual translation.
06/22/2026
Two ways to your own agent team in Claude Code and GitHub Copilot – adapt ready-made templates or build from scratch with /agents or /create-agent. Plus the trick for translating a Claude template into a Copilot-compliant variant via GitHub Copilot.
06/15/2026
A concrete walkthrough through all five phases of the agent team – from the first vague feature wish to a tested commit. Including the real handoffs, the checkpoints, and the moment where I had to correct the agents.
06/08/2026
Why a green build doesn't mean a feature works – and how a fifth, specialized agent turns the use case into an executable end-to-end test, closing the loop back to the original requirement.
06/01/2026
How an atomic battle plan and a focused Developer Agent close the gap between concept and code – and why Human-in-the-Loop works differently here than in the planning phase.
05/25/2026
How two specialized AI agents steer the planning process in the right direction through Human-in-the-Loop checkpoints – and why that matters more than pure speed gains.
05/18/2026
Why simple prompting and "vibe coding" fail on complex projects – and the four systemic problems that forced me to rethink my entire development workflow.
12/29/2025
My approach to improving code quality when using AI coding assistants through the use of copilot-instructions.md and AGENTS.md files.
12/18/2025
My experiences using AI in software development
12/14/2025
My experiences using AI in software development
12/02/2025
My experiences using AI in software development
06/11/2025
Installing and Using Ollama - A Guide for Local LLMs
06/11/2025
Installing and Using LM Studio - A Guide for Local LLMs
09/15/2024
In this tutorial, we'll extend our landing page with a multilingual content page that displays Markdown files using the nuxt/content module.
09/08/2024
This tutorial demonstrates how to implement a landing page using Nuxt, TailwindCSS, and i18n, including a hero section, feature section, and footer section with language switching.
09/06/2024
Experience Report - Usage of different AI Coding Assistance for private projects.
09/01/2024
This is a tutorial how to add interantionalization to a nuxt application and how to configure the module nuxt/i18n.
08/25/2024
This is a tutorial how to setup a new project with Nuxt and add modules like TailwindCSS, Nuxt/i18n and Nuxt Content v2.
05/11/2024
Are you ready to embark on an exciting journey where code meets creativity, and innovation fuels our passion for technology? If so, you're in the right place! 🚀