Automation & AI Agents

Workflows that think.
Agents that act.

I combine workflow automation, agentic AI and ML into systems that don't just save time — they make intelligent decisions, monitor, predict and act on your behalf.

Approach

Three building blocks. Infinite combinations.

The best automations emerge when workflow logic, agentic AI and ML work together — not as separate silos.

Workflow Automation

Time-scheduled and event-driven pipelines in n8n, Make and similar tools — triggering precisely when something happens or needs to happen.

The Core

Agentic AI

LLM agents that understand context, make decisions and execute actions — not just answering questions.

Machine Learning

ML models that find patterns, predict outcomes and give the system the ability to learn and improve over time.

Real-world examples

Four pipelines. Four problems solved.

01

News monitoring & content

n8n LLM LinkedIn Teams

Fetches news from selected sources every 6 hours, lets AI assess relevance and automatically writes LinkedIn posts ready for approval — delivered directly in the Teams channel.

Fetch news AI assesses relevance Write LinkedIn post Send to Teams
02

Intelligent mail triage with RAG

n8n RAG LLM

Monitors the info mailbox, analyses the content and automatically assigns it to the right person — including a draft reply based on the company's own knowledge base via RAG.

New mail received AI analyses content Assign to right person RAG generates draft reply
03

ML prediction of runner distance

Machine Learning Real-time Init-Together

ML model that continuously predicts how far a runner will get in 24 hours based on pace, heart rate and history — shown live so sponsors can follow along and bet on the runners.

Collect runner data ML predicts distance Show live for sponsors Update continuously
04

Self-developing feature pipeline

Agentic Claude Code Jira CI/CD

The client creates a feature in Jira — Claude Code develops it autonomously, deploys to a test server and waits for approval. When the client approves, it goes straight to production.

Feature in Jira Claude Code develops Deploy to test server Approve → prod

Technologies

What I work with

The tools that drive the automations — chosen for what solves the problem best, not what's trendy.

n8n & Make

Visual workflow platforms for orchestrating complex, event-driven pipelines without building everything from scratch.

LLM Agents

Claude, GPT-4 and open-source models in agentic setups — with tool use, function calling and multi-step reasoning.

RAG & Knowledge Bases

Retrieval-augmented generation that grounds AI answers in the company's own documents, databases and processes.

ML Models

Training and deployment of custom ML models for prediction, classification and anomaly detection in real time.

Get started

Do you have processes that can be automated?

Whether you want to save time on manual tasks, build intelligent agents or leverage your data with ML — let's talk about what makes sense.