GenAI product · Live
myAItinerary.com: an AI travel planner
A full-stack travel planner I built and deployed on my own, combining Google Maps, LLM-assisted itinerary editing via the OpenAI API, and a narrative user experience.
Open to industry roles after PhD
Machine Learning Engineer & Data Scientist.
PhD candidate in Neuro-Information Systems at the University of Freiburg with 5+ years of experience in machine learning, reinforcement learning, NLP, optimization, and data science. I enjoy building research-driven systems with practical impact.
- Python
- PyTorch
- Scikit-Learn
- SQL
- Git
- JavaScript
Selected work
Projects that show range and execution
Two projects you can try live, followed by research and applied ML work spanning deep learning, NLP, optimization, and geospatial analysis.
Reinforcement learning · Playable
Flip7: a PPO agent you can play against
I turned the card game Flip7 into a reinforcement learning environment, trained a PPO agent that beats a dynamic-programming opponent in 54.8 % of 10,000 games, and shipped a web app where you can play against it.
Lab experiment · GenAI
AI tone in itinerary planning
Does the tone of an AI assistant change how people use it? I built an AI travel planning platform and ran a lab experiment with 87 participants, and co-authored a related field study.
Lab experiment · NeuroIS
How people process deepfakes
I created political deepfakes with a generative AI pipeline and ran a lab experiment with 120 participants on how people detect them, combining eye tracking, ECG, and behavioral data.
Master's thesis · RL + OR
Solving the TSP with reinforcement learning
My master's thesis (grade 1.0) combined a PPO agent for route construction with simulated annealing and 3-Opt local search, finding routes up to about 25% shorter than a greedy baseline under strict time limits.
NLP · Empirical finance
Can Twitter sentiment improve volatility forecasts?
I collected almost one million tweets on six US stocks and added Doc2Vec and VADER sentiment signals to the HAR volatility model. For the two most-discussed stocks, forecast errors fell by 25–33%.
Deep learning · CV
Instagram image analysis with CNNs
I fine-tuned an EfficientNet-B3 network to score four content attributes in 4,070 Instagram images, more than halving the error of the best classical baseline, with Grad-CAM explanations.
Geospatial analysis
Classifying street networks of 58 cities
I grouped 58 cities worldwide by street-network type, using four structural metrics (grid perfectness, empty space, curvature, connectivity) computed with OSMnx and NetworkX.
Web scraping · Deep learning
Predicting rent from kitchen and bathroom photos
I built a scraper that collected 6,631 Berlin listings with photos and trained a two-branch CNN on one kitchen and one bathroom photo per apartment to predict rent per m².
Web development
Digital wedding invitation & gallery
I designed and built an invitation and media gallery site end to end, a small project in front-end craft and visual design.
And more
Explore the code
Additional projects, experiments, and the source behind this work.
Research output
Publications
2025 · Preprint
Persuasive or Neutral? A Field Experiment on Generative AI in Online Travel Planning
Jirpongopas, L., Lutz, B., Ebner, J., Vahidov, R., Neumann, D. arXiv:2509.14259.
2024 · Research-in-progress · NeuroIS Retreat, Vienna
Deep Fakes and Political Attitudes: An Analysis of Confirmation Bias and Cognitive Dissonance Using Neurophysiological Measurements
Ebner, J., Lutz, B., Neumann, D. In: Information Systems and Neuroscience, LNISO vol. 66, pp. 299–311. Springer.
Background
Experience and education
Professional experience
08/2022 – present
Research assistant / PhD candidate
Chair of Information Systems Research, University of Freiburg
Research in NeuroIS, teaching in Optimization & Simulation.
Built a generative-AI deepfake pipeline and UIs for lab studies.
Responsible for the chair's web presence and technical support.
04/2025 – present
Lecturer at DHBW Lörrach
Business Administration (Industrial Management), Baden-Württemberg Cooperative State University (DHBW), Lörrach
Courses: Introduction to Economics and Microeconomics, Advanced Microeconomics.
12/2019 – 07/2022
Student research assistant
Chair of Information Systems Research
Supported research with Python, contributed to teaching, and maintained web infrastructure.
03/2019 – 03/2021
Student research assistant
Chair of Statistics and Econometrics
Created lecture slides and exams in LaTeX, organized teaching materials, and managed website responsibilities.
Education
08/2022 – present
PhD candidate, Neuro-Information Systems
University of Freiburg
Research at the intersection of machine learning, behavioral experimentation, and physiological measurement.
10/2019 – 05/2022
M.Sc. Economics (final grade 1.2)
University of Freiburg — Business Analytics and Empirical Economics
Master's thesis (1.0): Reinforcement Learning for Combinatorial Optimization → case study
10/2016 – 09/2019
B.Sc. Economics (final grade 1.4)
University of Freiburg
Bachelor's thesis (1.0): Machine Learning in Real Estate — rent prediction with NLP and gradient boosting.
Skills
Programming
- Python
- JavaScript
- R
- HTML/CSS
- PHP
Data Science
- NumPy
- Pandas
- SciPy
- Scikit-Learn
- XGBoost
Deep Learning
- PyTorch
- TensorFlow
- Keras
RL
- Gymnasium
- Stable Baselines3
NLP
- Word2Vec
- Doc2Vec
- NLTK
- VaderSentiment
- Transformers
Geospatial Analysis
- Geopandas
- OSMnx
- NetworkX
GenAI / LLMs
- OpenAI API
- Prompt Engineering
- LLM Integration
Engineering
- Git
- SQL
- Flask
- REST APIs
- Cloud Deployment
Experimentation
- MNE
- NeuroKit2
- oTree
- APIs
Languages
- German (native)
- English (C1)
- Spanish (A1)
Interests
- Swimming/DLRG
- Powerlifting
- Alpine Hiking
- Board and strategy games
Contact
Ready for the next chapter
Best-fit opportunities
I'm looking to join a team building applied AI systems, bringing research rigor to work that moves from experimentation to real product impact.