HAW Hamburg · Humboldt University · 13+ supervised theses HAW Hamburg · Humboldt-Universität · 13+ betreute Arbeiten

Teaching Lehre

I teach where industrial AI engineering meets academic rigor — and supervise theses that publish. Ich lehre an der Schnittstelle von industrieller KI-Praxis und akademischem Anspruch — und betreue Abschlussarbeiten, die publiziert werden.

Master Summer Term 2024 & 2025Sommersemester 2024 & 2025

Advanced Programming

Master Media Informatics · HAW Hamburg · Finkenau 35, 22081 Hamburg Master Medieninformatik · HAW Hamburg · Finkenau 35, 22081 Hamburg

Co-taught with Prof. Dr. Larissa Putzar. Two contact hours per week. Students deepen the theoretical foundations of modern AI in lectures and apply advanced engineering concepts in tutorials and project work — building, probing and analyzing real machine-learning systems. Gemeinsam mit Prof. Dr. Larissa Putzar. Zwei Kontaktstunden pro Woche. Die Studierenden vertiefen die theoretischen Grundlagen moderner KI in Vorlesungen und wenden fortgeschrittene Engineering-Konzepte in Übungen und Projekten an — sie bauen, prüfen und analysieren echte Machine-Learning-Systeme.

TopicsThemen

  • Robustness probing of classical and deep modelsRobustness-Probing klassischer und tiefer Modelle
  • Explainability methods (SHAP, attention analysis, counterfactuals)Erklärbarkeitsmethoden (SHAP, Attention-Analyse, Counterfactuals)
  • Large language models in production pipelinesLarge Language Models in Produktiv-Pipelines
  • Practical project: build & audit a small ML systemPraxisprojekt: Aufbau & Audit eines kleinen ML-Systems

Interactive demos used in seminars and student exercises. They make abstract concepts — robustness, drift, explainability — tangible in the browser. Interaktive Demos für Seminare und studentische Übungen. Sie machen abstrakte Konzepte — Robustheit, Drift, Erklärbarkeit — direkt im Browser greifbar.

interactive tutorial fundamentals

Perceptron TutorialPerceptron-Tutorial

Hands-on walkthrough of the perceptron — the building block of neural networks. Tweak weights and bias, watch the decision boundary move in real time.Praktischer Einstieg in das Perzeptron — den Grundbaustein neuronaler Netze. Gewichte und Bias verändern, die Entscheidungsgrenze bewegt sich in Echtzeit mit.

prototypes/perceptron.html ↗
explainability notebook

XAI PlaygroundXAI-Playground

Notebook companion: SHAP, attention overlays and counterfactuals on a shared text-classification benchmark.Notebook-Begleiter: SHAP, Attention-Overlays und Counterfactuals auf einem gemeinsamen Text-Klassifikations-Benchmark.

comingfolgt
llms eu-ai-act

Doc-Compliance SandboxDoc-Compliance Sandbox

Companion to the IJCNN 2025 paper: drop a documentation PDF, see article-level checks in real time.Begleiter zum IJCNN-2025-Paper: PDF einfügen, Artikel-genaue Prüfung in Echtzeit.

comingfolgt

Over a decade I've supervised more than 13 Bachelor's, Master's and doctoral theses across AI, machine learning and audit technology. Several have led to peer-reviewed publications. Über ein Jahrzehnt habe ich mehr als 13 Bachelor-, Master- und Doktorarbeiten in KI, Machine Learning und Audit-Technologie betreut. Mehrere führten zu begutachteten Publikationen.

2025
  1. PhD ongoing · Computer Science
    Enhancing Transparency of AI Decision-Making Algorithms by Leveraging Large Language Models
    Q. Tran · since 2024
  2. Bachelor
    Generative Artificial Intelligence in Auditing: Overcoming Barriers to Comprehensive Audits and Reducing Detection Risk
    H. Neupert
2024
  1. Bachelor · Financial Mgmt., Accounting & Taxation
    Quality Assurance of Generative Artificial Intelligence in Auditing: Conceptual and Regulatory Approaches
    N. Wetzel
  2. LL.B.
    Trustworthy AI in Auditing: The Role of General-Purpose AI Models and Regulatory Requirements under the EU AI Act Using ChatPwC as a Case Study
    I. Mayer
2023
  1. Master · Computer Science
    The Impact of Anonymization on the Explainability of Machine Learning Models
    K. Shpileuskaya
  2. Bachelor · MBA
    Requirements for the Implementation of Generative AI in Audit Reporting
    S. Eibl
  3. Bachelor · Data Science
    Analyzing Business Implications of Large Language Models: A Study of the Models
    J. Salg
2022
  1. Master · Business Informatics
    Assessing the Robustness of a Machine Learning-Based Information Extraction System in Audit Processes at PwC
    Q. Tran
  2. Master · Digital Reality
    Comparative Study of Generative Adversarial Network Architectures for Time Series Forecasting
    V. Schnorr
2021
  1. Master · Applied Data Science
    Explainable Artificial Intelligence for Creditworthiness Assessment
    M. L. Struckmann
  2. Master · Applied Mathematics
    Analysis of Numerical Methods for Arbitrage Optimization: Monte Carlo vs. Quantum-Inspired Simulated Annealing
    J. S. Dürr
2020
  1. Master · Financial Mathematics
    Robustness Analysis of Machine Learning Methods in Automated Trading Systems
    P. Stein
  2. Bachelor · Business Informatics
    Robustness Analysis of Machine Learning Methods for Document Classification
    Q. Tran
2019
  1. Bachelor · Business Informatics
    Ensuring Audit Trails through a Prototype Implementation Using Blockchain Technology
    T. Oegel

Conference talks at IJCNN, ACII and PETRA Konferenzvorträge auf IJCNN, ACII und PETRA

I regularly present joint work with my collaborators at HAW Hamburg, and run internal workshops at PwC on AI auditing, EU AI Act readiness and LLM evaluation. Ich präsentiere regelmäßig gemeinsame Arbeit mit den Kolleg:innen der HAW Hamburg und halte interne PwC-Workshops zu KI-Audit, EU-AI-Act-Readiness und LLM-Evaluation.

Invite a talkVortrag anfragen