Alexander Seipel
EN DE Standard Lite Contact
Menu
Version Standard Lite

AI and software engineering

Prague, Czechia 50.08° N 14.44° E

Alexander Seipel

I work out where AI pays off.
Then I build it.

I have built software for ten years, from a startup where I was co-owner and CTO to Commerzbank’s online banking. The advice and the code come from the same person.

alex@seipel.me CV (PDF) (opens in a new tab) LinkedIn (opens in a new tab)

Open for new projects

01 / System simulation
years building software
10
industries
6
years as co-owner and CTO
4½

Where I have worked

  • Commerzbank
  • MaibornWolff
  • Brandl Nutrition
  • Immowelt
01 Consulting and delivery

Services

01 / 03

AI consulting

I find the work in your company where a model pays for itself, and I tell you where it does not. At Commerzbank I already help the engineering teams adopt AI tools.

Assessment
I sit with the people who do the work and see where their time goes. Sometimes the answer is that no model belongs anywhere near the task. You get that in writing too.
Tool selection
You can buy one tool, connect several, or build your own. I tell you which option fits and what it costs, both to run each month and to leave later.
Enablement
A tool nobody opens is just a subscription. I run the training sessions and write the guidelines your team follows.

02 / 03

AI software

I design and build the AI features that are specific to your product. Off-the-shelf tools only cover what every company has in common.

A better interface
A chat box asks people to describe work they already know how to do. A screen built for the task is quicker, and it leaves fewer places for a wrong answer to hide.
Model selection
Models differ in price by a factor of ten or more, and their quality on your task differs just as much. I test them on your data and use the expensive one only where it pays for the difference.
Your own documents
An assistant that answers from your own files first has to find the right passage. That search step, called retrieval, is where these systems usually fail. It is also where I started: my bachelor’s thesis evaluated a learning-to-rank search prototype.
The wrong tool
A model is often the wrong tool. If a database query or a better form solves the problem for less, I’ll tell you so, even though it means I build less.

03 / 03

Software development

Most of my ten years came before this wave of models. I build web applications, APIs, online shops, and the internal tool everybody needs and nobody wants to own.

Prototype first
You get something to click on early, while a change of direction is still cheap. Then we turn it into a product your team can maintain.
Development with AI
In my own work, AI agents now write most of the code. Claude Code and Codex are open on my screen every day. The real work has moved to the specification and the review, and I do both myself.
Architecture
I work with AWS or GCP, Terraform, Docker and CI/CD. I once moved a whole shop from a single rented server to auto-scaling AWS with Aurora, CloudFront and a WAF, all managed in Terraform.
E-commerce
I replaced Magento with headless Shopify: Shopify ran the checkout behind a Gatsby storefront I built, with subscriptions and automated discount campaigns. I also build on the Stripe API.
Accessibility
Since June 2025 the European Accessibility Act has applied across the EU, in Germany as the BFSG. A bank audits WCAG conformance instead of assuming it, and I build to that standard. I also bring older interfaces up to it.
02 2016 to now

Experience

I build software that has to hold up in production.

For four and a half years I was co-owner and CTO of Brandl Nutrition, and its only engineer. I owned the recipe engine, the shop, the fulfilment down to the printed label, and the AWS account under all of it.

Before that I trained neural networks at Immowelt, part of the Axel Springer group, and put them into production. In 2023 I joined the consultancy MaibornWolff in Berlin and worked for three clients in one year. Then I moved to Prague, and since September 2024 I have been building frontends for Commerzbank.

Outside work I lift weights and plan my meals in a spreadsheet, down to the gram. That habit took me to a nutrition startup in the first place. It is also how I build software: measure first, then decide.

Alexander Seipel on the Vltava embankment in Prague
Prague
Based in
Prague, Czechia · CET
Remote
Worldwide
Now
Commerzbank AG, Engineer
Languages
German, English

Positions

  1. Prague · online banking and advisory

    • Transaction Overview is the transaction list millions of customers see after they log in. I am the only frontend engineer on it, and I moved its toolchain from React Scripts to Vite.
    • For a new advisory platform I architected the React frontend from the first commit and built several of its modules alone. The platform has AI features, and the details are confidential.
    • Accessibility is part of the definition of done: every feature meets WCAG before it ships.
    • I write the OpenAPI specifications with the backend teams, and I onboard new developers.
    • I help the engineering teams adopt AI tools and train colleagues to use them.

  2. Berlin · three clients

    • A sustainability startup: the frontend of an import tool for supplier and farm data, in React with AG Grid and GraphQL. It was my first direct client work, so I ran the requirements workshops before I built anything.
    • A testing and certification group: a proof of concept for a token-based design system, from Figma tokens synced through Git to web components in StencilJS. Wrapper packages let the React, Angular and Vue teams pull the same version.
    • A medical technology company: new features and fixes on a therapy platform for clinics worldwide. I had never written Angular before. Within days I was productive, and I mentored the junior developer on the team.

  3. Nuremberg · sole engineer

    • I built the engine that turned a customer questionnaire into an individual supplement recipe. Behind it, a Django backend produced the production sheet, the back label, the delivery note and the DHL label.
    • Brandl recorded a pitch for “Die Höhle der Löwen”, the German Dragons’ Den, and it could air on two weeks’ notice. I moved the platform from one rented server to auto-scaling AWS, managed in Terraform, so the shop would stay up. The episode never aired.
    • Magento went out and headless Shopify came in, behind a Gatsby storefront I built. The marketing stack was mine too: analytics, A/B tests, newsletter automation, and the Python scripts that analysed the results.
    • As co-owner I had a say in the range as well. The plain protein I pushed for became the best seller.
    • From August 2023 I ran all of it alongside my work at MaibornWolff, until I left in June 2024.

  4. Nuremberg · big data team

    • I started in the big data team as an intern in October 2016 and joined full-time in June 2017.
    • I trained machine learning models and put them into production: price estimation, image classification, and data extraction from photos of energy certificates. I used TensorFlow, Keras, PyTorch and scikit-learn.
    • To fight fraud on immowelt.de, I built a mobile TAN flow, connected external identity verification providers, and wrote detection services in Python with scikit-learn.
    • “Markt und Preise” is a public tool where anyone can chart German rent and purchase prices by the criteria they choose. I was its architect and full-stack developer, with C# and .NET on ClickHouse.
    • I worked across SQL Server, Vertica, Couchbase, MongoDB and ClickHouse. For MongoDB I wrote a real-time library that started a downstream pipeline whenever the data changed.

  5. Nuremberg

    • My first paid job, during my studies: search engine optimisation for client websites, plus online and social media marketing.

  6. Regensburg · information retrieval

    • I built a search prototype on Apache Solr with learning-to-rank.
    • I analysed the Twitter log data behind the experiments in R, and the write-up became my bachelor’s thesis.

From my employment references

“Mr Seipel’s performance earned our full recognition at all times and in every respect.”
Employment reference · Immowelt AG · translated
“He found excellent solutions for all arising problems without exception.”
Employment reference · MaibornWolff GmbH

On the side

I built RedeStudio, Behandlung nebenan and NewBuildTracker alone in September 2026.

Behandlung nebenan (opens in a new tab)

Behandlung nebenan shows patients with German statutory insurance what their dental treatment costs at German-speaking clinics in Czechia. They upload the plan from their dentist and get an estimate within 24 hours.

NewBuildTracker (opens in a new tab)

NewBuildTracker follows the prices of new-build flats in Prague, like a flight-price tracker. Every night a model reads the developers’ price lists and records each change.

impflotse.de (opens in a new tab)

Patients book vaccination appointments at pharmacies online. I’m its only developer and have worked AI-assisted since the first commit. I also handle the GDPR side, from legal logging to automated data disclosure.

Affiliate site

Earlier I built an affiliate site on Gatsby with a friend, and it earned money. The search engine optimisation I had learned in my first job went into it.

03 What I build with

Stack

AI

  • LLM APIs
  • Claude
  • GPT
  • Agents
  • MCP
  • RAG
  • Vector search
  • Evaluation
  • Open models
  • TensorFlow
  • scikit-learn

Languages

  • TypeScript
  • JavaScript
  • Python
  • Java
  • C#
  • SQL

Frontend

  • React
  • Next.js
  • Angular
  • Vue
  • GraphQL
  • Web components

Backend

  • Node.js
  • Django
  • FastAPI
  • .NET
  • REST
  • OpenAPI
  • Stripe
  • Serverless

Infrastructure

  • AWS
  • GCP
  • Terraform
  • Docker
  • CI/CD
  • PostgreSQL
  • ClickHouse
  • MongoDB
Education
B.A., University of Regensburg Media Computer Science and Information Science, 2013 to 2017. My thesis evaluated a learning-to-rank search prototype.
Spoken languages
German: first language English: level C1
Clients
Startups, mid-sized companies, agencies, large enterprises On my own, or as part of your existing team.
04 How I work

Approach

No slide deck.
No army of consultants.
One engineer who builds it.

01 Scope
Before I start, we agree what is in and what is out. Short cycles keep the cost visible and make it cheap to stop.
02 Evidence
I test every AI claim before it gets a budget, including my own. If the numbers do not hold up, you hear it from me first.
03 Ownership
The code, the infrastructure and the documentation are yours, and nothing runs on an account that only I can reach.
04 Progress
You see running software rather than a status report about running software.
05 Leadership
If your team has no technical lead, I take on the role. I have done it before. For four and a half years I was responsible for all of a company’s technology, with nobody to hand it to.

Engagement

  • Remote from Prague, CET
  • German or English
  • Day rate or fixed price, agreed per project
  • From a short assessment to a full build
05 Prague · CET

Contact

Tell me about your project.

alex@seipel.me

I usually answer within a day.

CV (PDF) (opens in a new tab) LinkedIn (opens in a new tab)
Alexander Seipel on a rooftop in Berlin