Skip to content
WD
PL EN

AI Automation Engineer

Wojciech Dekowski

I turn repetitive business processes into intelligent automated systems. I design architecture, integrate APIs, and deploy AI models to scale what companies can do.

Talk to AI View projects
wojciech@ai-clone ~ en

About

I combine software engineering with a deep understanding of business processes. I specialize in workflow orchestration (n8n, Python) and deploying language models (LLMs, RAG, AI Agents) to solve real problems. My goal is simple: make technology work for people.

Experience

Production Process Optimization Specialist

Manufacturing environment · Industry experience

  • Ongoing analysis and optimization of production technology chains.
  • Real-time problem solving in high-rigor manufacturing environments.
  • That experience built my foundation in systems thinking — today I apply the same optimization principles to digital architecture and API networks.

Projects

Architecture and business outcomes — not just code.

Sales Automation System

n8n Webhooks REST API
Problem
Manual retyping of order data and fulfillment delays caused errors and bottlenecks in customer operations.
Architecture
n8n orchestration with inbound webhooks and REST API integrations across sales and warehouse systems.
Solution
A zero-touch pipeline from order intake to confirmation — data flows between systems without manual intervention.
Business impact
Faster order handling and higher team throughput by removing repetitive operational work.

AI Automation Platform

Python Docker Portainer Gemini API n8n
Problem
Scattered scripts and no shared runtime made automation hard to scale or ship to production.
Architecture
Python + Docker + Portainer as the runtime layer, Gemini API for language models, n8n for workflow orchestration.
Solution
A unified platform to run AI agents and APIs with container control and production-ready deployment paths.
Business impact
A scalable agent orchestration environment — from prototype to production without ad-hoc script chaos.

AutoDocs AI

RAG Vector DB Gemini OpenAI API
Problem
Reviewing hundreds of pages of technical and business documentation took hours and slowed decisions.
Architecture
RAG pipeline: document chunking, embeddings, a vector database, and queries via Gemini / OpenAI API.
Solution
Semantic search and information extraction from documentation in natural language.
Business impact
Near-instant access to documentation facts instead of manually scanning hundreds of pages.

KartonPak

Python Business Process Digital Workflow
Problem
No digital data flow on the shop floor — information was lost between paper and systems.
Architecture
An application supporting a real production process by digitizing information flow on the floor.
Solution
A working tool in a demanding physical business environment — a bridge between operations and data.
Business impact
Demonstrates understanding of physical company processes and shipping solutions where lab-perfect conditions do not exist.

WatchDog

n8n Docker Webhooks Healthcheck
Problem
Many services and portfolio demos run in containers — a failure stays silent until someone manually checks logs.
Architecture
n8n loop: URL healthcheck → IF down → Docker container restart → notify (Discord) + POST live status to the portfolio.
Solution
An automated uptime guardian: detects down services, brings them back up, and reports live status on the site.
Business impact
Less manual babysitting of the server, faster downtime response, and tangible proof of ops/automation skills.

n8n · live

WatchDog — live monitoring

An n8n automation checks whether projects are up. If something fails, it restarts the Docker container and sends an alert. Statuses below stream live from the webhook.

Live status
Waiting for the first n8n payload…

Waiting for the first n8n payload…

Talk to my digital clone

An AI assistant answers questions about projects, stack, and collaboration — no hallucinations, facts only.

wojciech@ai-clone ~ en

AI Lab

Short case studies on automation, prompt engineering, and RAG.

Read more →

WatchDog: n8n watches containers and restarts them

Healthcheck → IF down → Docker restart → Discord + portfolio status. An uptime automation case study.

How I wrote a prompt that cut hallucinations by ~80%

Hard system rails, a knowledge JSON, and a ban on guessing — a practical pattern for a portfolio AI assistant.

My n8n schema for email analysis

Webhook → classification → field extraction → CRM/Slack routing. A simple pipeline that kills manual inbox triage.

Contact

Let's talk about automation that actually saves time and money.

Send a message

miltion6@gmail.com