Research
🔬 I am an Applied AI Researcher at BMW Group in the GenAIM team, working on agentic AI systems for procurement and suppliers.
My research sits at the intersection of causal discovery, sequence modeling, large language models, and agentic architectures — building AI systems that don’t just predict, but reason about why and what if.
Previously, I have submitted my PhD in Data Science at the University of Augsburg advised by Prof. Dr. Rainer Lienhart, where I developed sequence models and causal discovery methods for high-dimensional event sequences in vehicle diagnostics.
Current research directions:
- Causal foundation models for tabular and structured enterprise data
- Multi-modal context management for long-horizon agentic conversations
- Agentic tool-selection and planning under uncertainty
⭐️ Open Source:
- 📦
seq2cause: A Python package for population and sample-level causal discovery in high-dimensional event sequences via neural autoregressive density estimation.
💻 Stack: PyTorch, Transformers, LLMs, Causal Discovery, Graph Representation Learning.
News
- June 2026: Paper accepted to ICML 2026 workshop (SPIGM) — Your Autoregressive Model Already Reveals the Causal Graph.
- June 2026: 🏅 Awarded ICML 2026 Gold Reviewer distinction — signed letter.
- June 2026: Reviewer for NeurIPS 2026.
- March 2026: Two papers accepted to ICLR 2026 workshops — presenting work on multimodal vehicle diagnostics and neuro-symbolic LLMs.
- Sept 2025: Two papers accepted to NeurIPS 2025 workshops (SPIGM & CauScien), San Diego.
- Feb 2025: Presented at AAAI 2025 — event sensory data and language models for vehicle diagnostics.
Selected Publications
(Full list on Google Scholar)
| Paper | Venue | Authors |
|---|---|---|
| Your Autoregressive Model Already Reveals the Causal Graph | SPIGM @ ICML 2026 | Hugo Math, Rainer Lienhart |
| Context-Informed Sequence Classification: A Multimodal Approach to Vehicle Diagnostics | TSALM @ ICLR 2026 | Hugo Math, Rainer Lienhart |
| Neuro-Symbolic Rule Discovery: Empowering LLMs with Causality for Vehicle Diagnostics | Logical Reasoning @ ICLR 2026 | Hugo Math, Julian Lorenz, Rainer Lienhart |
| One-Shot Multi-Label Causal Discovery in High-Dimensional Event Sequences | CauScien @ NeurIPS 2025 | Hugo Math, Robin Schön, Rainer Lienhart |
| Harnessing Event Sensory Data for Error Pattern Prediction in Vehicles: A Language Model Approach | AAAI 2025 | Hugo Math, Rainer Lienhart, Robin Schön |
Beyond Research
- 💡 Industry: Freelance computer vision engineer — automated semiconductor wafer inspection systems.
- 🇸🇪 Exchange: Chalmers University of Technology, Gothenburg, Sweden.
I’m always happy to chat about research, causal AI, or agentic systems — reach out on LinkedIn or by email.