I turn
technology

into impact

I advise on AI and build the systems myself, from strategy to running operations.

Worked with

AllianzSiemensHeinrich-Böll-StiftungUniversität BernMANHochschule MünchenLufthansaHasso-Plattner-InstitutNokiaTactical TechEvonikStrascheg Center for Entrepreneurship (SCE)KertosVitolusAllianzSiemensHeinrich-Böll-StiftungUniversität BernMANHochschule MünchenLufthansaHasso-Plattner-InstitutNokiaTactical TechEvonikStrascheg Center for Entrepreneurship (SCE)KertosVitolus

Starting point

This is how most first calls begin.

In all four cases, what is missing is an independent judgement that looks at technology, data protection and business together.

Ways to work with me
  • “The board expects an AI strategy. I have three proposals on the table, and each one tells a different story.”
  • “Half the team has been using ChatGPT for ages. Rules, approvals, data protection: all unresolved.”
  • “Our data can’t simply go into a US cloud. So what is still possible?”
  • “The pilot looked good. It never went live.”

Services

Three ways to work with me

Review: Assess an AI initiative

Need a sound assessment before you make an AI decision?

I review your initiative before you invest. The review has a fixed price and a clear end. Independently and within a clearly defined scope, I examine the architecture, the measurable quality criteria (evals) and the security of your AI initiative. You get a written result with prioritised recommendations: what is possible, what is worth it and what is better left alone.

Fixed price · Clear end · Written result

AI review

2 weeks · Fixed price

Build: Build an AI system

Your AI pilot convinces in the demo but never reaches production?

I bring pilots into production. This is where many projects stall: legacy systems, patchy data, unclear ownership. Prototype, MVP, AI platform or agent system: connected to your existing systems (MCP integrations) and continuously checked for quality (evals). You get a defined outcome that I deliver independently with my own tools, including the handover to your team.

Fixed-price delivery · Defined scope

AI-native setup

Fixed price · Your AI operating system

Use-case lighthouse

6 weeks · A running system

Interim: Technical leadership for a set period

You need technical leadership for several months, but no full-time CTO?

I take on technical responsibility. As interim or fractional CTO, I make architecture decisions and set what gets built in which order, from the first idea to running operations. The scope is a fixed monthly allotment with agreed milestones and a clearly defined area of responsibility. The mandate is time-limited, and responsibility passes to your team step by step.

Monthly allotment · Clearly defined scope · Time-limited

Interim mandate

3 to 6 months · Monthly allotment

Fractional CTO

6 months or longer · Monthly allotment

Fixed-price packages available I do the work myself At most three mandates at a time

Selected projects

AI platform Allianz SE

Chatbots that service teams build themselves

An AI-supported no-code platform that lets service teams build their own chatbots. Thousands of requests are handled faster, and analytics make service quality measurable.

Platform Heinrich-Böll-Stiftung

One platform for a worldwide network

Hundreds of partners around the world work together on one shared platform. Open source, hosted in the EU, GDPR-compliant and with no data flowing to Big Tech.

Architecture Kertos

AI platform from scratch

Architecture and agent system for a compliance platform, from the first line of code to LLM observability in production. MCP integrations, evaluation pipelines, Datadog.

Engineering Vitolus Services

AI for better diagnostics

From idea to running product: custom AI models for image recognition and report generation in medical technology. Built from zero as lead engineer.

Analytics Hasso-Plattner-Institut

Making learning visible

An analytics platform that evaluates the learning behaviour of hundreds of thousands of users in real time and turns it into personal recommendations.

Data architecture Siemens AG

Supply chain data in seconds, not minutes

Redesigned the data architecture behind a virtual factory model, backed by benchmarks and a proof of concept. The central supply chain view now loads in seconds instead of minutes, connected to SAP.

What clients say

Translated from German.

“Simon designed the architecture around our compliance agent KAIA and brought it to paying customers. He was instrumental in establishing evals and made the quality of our AI measurable. Simon takes on far-reaching responsibility, brings the team along and ensures the highest level of security and reliability, even on the most complex tasks.”

AP

Alexander Prams

CTO & Co-Founder, Kertos

“As our fractional CTO, Simon designed and built our first MVP with us, from the data model to the first tests with care workers. He brings structure to our ideas and helps us build the right things in the right order. We now develop our AI tools ourselves, with Simon as a sparring partner. When we get stuck, he helps us move on, and he makes sure security and data protection are in order.”

MD

Moritz Dillmann

Co-Founder & CEO, Workinger

My approach

Problem first. System second.

That is why every project starts with listening. Technology comes later.

01

Understand

Often the solution is settled before the problem has a name. So I first clarify what should change in a measurable way and how you will recognise success. Without a clear goal, a project gets expensive long before it gets smart.

02

Assess

Which technology fits, and which does not? Which base model (foundation model)? Connect your data or train on it (RAG vs. fine-tuning)? Your own servers or the cloud? Not every problem needs AI.

03

Build

AI can build a prototype in hours today. Most of the work comes afterwards: connecting it to your data and systems and running it reliably. That part I build myself.

04

Operate

In operation, I measure quality continuously (evals), make visible what the system does (LLM observability) and improve it on that basis. An AI system is never finished.

05

Hand over

From day one, I work so that your people can follow along: with documentation, an eval suite (automated quality tests) and code reviews that explain the reasoning. That way your team can run and develop the system itself whenever you want.

What I don’t do

No forever POCs.

If there is no reliable system after 6 weeks, the use case is not the problem.

No lock-in architectures.

Your own team has to be able to keep building without me.

No “live without evals”.

Without evaluation, nobody knows whether an AI system gives the right answers in production.

Compliance is architecture

GDPR and the EU AI Act built into the design. Data protection and documentation from the first sketch.

Data stays where it is

On-premise, sovereign AI, EU hosting down to the hardware.

Built in the open

Open source where possible, no vendor lock-in, no black box.

Simon Liedtke on stage at re:publica
I only advise on what I can build and verify myself.

Simon Liedtke Munich 20+ years in software 5+ years full-time in AI LinkedIn

I have worked in startups, mid-sized companies and large German corporations (DAX). I know all three from operational responsibility as engineer, architect and co-founder.

Today I work as an independent consultant with a lead engineer background. Architecture and code stay in one hand, and I take responsibility for systems all the way into operation. Current focus: AI architecture, agent systems and production AI, from the first line of code to measurable operation (LLM observability).

What I believe: AI amplifies human judgement.

Before we start

Frequently asked questions

How is data handled?

By default, there is no training on client data and no logging to external services without your explicit approval. For sensitive use cases, I use an on-premise or air-gapped setup on your side. An NDA comes first, as a matter of course.

And the EU AI Act?

For me, compliance is part of the architecture from the start: GDPR-compliant processing and EU hosting are standard in what I build, and the obligations from the EU AI Act (such as AI literacy in the team, Article 4) feed into recommendations and training. I don’t provide legal advice. I build so that your lawyers have little to object to.

How is intellectual property protected?

Code, architecture and methodology stay with you. There is no reuse in other client projects without your approval. Because I work on my own, clients stay separate by design: there is no shared team working on similar use cases in parallel.

What happens after the project ends?

Architecture and code stay with you. Your team has to be able to keep building without me. The handover, including documentation, an eval suite (automated quality tests) and onboarding notes, is part of the delivery.

We already use ChatGPT or Copilot. Why do we need advice?

What is usually missing: rules, shared ways of working, quality checks and a clear view of which use cases are worth it. The audit sorts that out, building on what you already have.

Why one person instead of an agency?

You get senior experience throughout the project, from the first call to the delivery. There is no handover to juniors and no standard template. I take on at most three mandates at a time, and your questions are answered by the person who builds.

Can you take over the whole implementation?

Yes, I can implement all of it, from the architecture to the running system. What I avoid is a full-service package that leaves you permanently dependent on a provider. I build with your team, as much or as little as you like, and make sure you can keep building without me.

How soon can we start?

A first call is usually possible this week, an audit usually within four weeks. Interim mandates and fractional CTO engagements can start at short notice. We’ll settle the timing in the first call.

Remote or on site?

Both. Workshops and board sessions on site with you, architecture work and code mostly remote, which saves time on both sides. In the Munich area, a hybrid setup is easy.

Can you promise results?

I don’t promise returns. Whether AI pays off depends on the use case, and nobody can seriously promise “30 to 50% savings” up front. What I commit to is the scope: the audit ends with a clear assessment after two weeks, and the lighthouse aims at a running system within six weeks. My recommendations are honest, including when the answer is no. And if the roadmap is not concrete enough for your team, I refine it at no extra charge.

What if it turns out AI won’t help us much?

Then the result says so, along with what will help instead. Some problems need better data or clearer processes, and no AI at all.

What does it cost?

I don’t publish list prices, because every situation is different. After the first call you get a fixed price, with no estimate and no “starting from” offer. A concrete quote usually follows within five working days of the briefing.

Contact

Email me at info@slworks.de or book a free intro call right away.

Book an intro call

Loads the Cal.com booking calendar (USA). This transmits your IP address to Cal.com. More in the privacy policy (German).