Munich🥨NLPxXpeng: Synthetic Data Generation

​Munich🥨NLP x XPeng: Synthetic Data Generation & Multilingual LLM Evaluation

​Hello Munich NLP Enthusiasts!

​We’re thrilled to announce our next meetup, hosted in collaboration with XPeng at their office (Weimamer Str. 32, 80807 Munich) on Wednesday, June 10, 2026! Join us for an evening of cutting-edge research, industry insights, and networking at the intersection of synthetic data generation and multilingual LLM evaluation.

​This event features two compelling talks:

• ​Florent Duême & Galina Lavrenteva (XPeng) will explore how synthetic dialogue generation and generative audio methods can create large-scale training and evaluation data for voice assistants — from realistic multi-turn conversations to augmented acoustic environments.

• ​Raoyuan Zhao (LMU Munich, MaiNLP Lab) will dive into synthetic data for LLM evaluation, addressing dynamic, scalable, and multilingual assessment challenges.

​📅 Agenda

​18:00 | Doors open + 🍕 food & drinks (pizza, juice, schorle, soda)
18:45 – 19:00 | Intro (XPeng, MunichNLP)
19:00 – 19:40 | Talk 1: "Synthetic Dialogue Generation & Audio Augmentation for Voice Assistants" Speakers: Florent Duême & Galina Lavrenteva (XPeng AI Model Team)

​Abstract:
Discover how XPeng generates large-scale synthetic dialogues and augments them with realistic acoustic environments for voice assistant training and evaluation. Built on SDialog (arXiv:2506.10622), this talk covers the full pipeline: LLM-orchestrated persona-driven text generation, controlled evaluation of dialogue quality, and generative audio methods for simulating background noise and spatial sound conditions. The result is a scalable framework improving both NLU and ASR robustness.

​About the Speakers: Florent Duême focuses on synthetic text data generation and evaluation—designing persona-driven dialogue orchestration, scenario scripting, and automated quality metrics. Galina Lavrenteva specializes in audio augmentation, leveraging generative audio to simulate realistic acoustic environments and noise conditions. Both are part of XPeng’s Munich AI Model Team, using the SDialog framework (Burdisso et al., 2025) to bridge generative AI with real-world voice assistant challenges.

​19:45 – 19:50 | Short break

​19:50 – 20:30 | Talk 2: "Synthetic Data for LLM Evaluation: Toward Dynamic, Scalable, and Multilingual Assessment" Speaker: Raoyuan Zhao (LMU Munich, MaiNLP Lab)
Abstract:
Static, English-centric benchmarks struggle to keep pace with LLM advancements. Raoyuan will present recent work on synthetic data for controllable, adaptive evaluation, including perturbations to probe model robustness (e.g., typographical variation) and methods to reduce data contamination. She’ll also discuss reliability, diversity, and efficiency in evaluation frameworks.

​About the Speaker:
Raoyuan Zhao is a PhD student at LMU Munich’s MaiNLP Lab, supervised by Dr. Michael A. Hedderich. Her research focuses on LLM evaluation, synthetic data, reinforcement learning, and multilinguality, with a commitment to advancing scalable and inclusive AI assessment.

​20:30 – 21:00 | Networking & drinks

​🔍 Event Details

​📍 Location: XPeng Munich Office, Weimamer Str. 32, 80807 Munich
📅 Date: Tuesday, June 10, 2026
⏰ Time: 18:00 – 21:00 (talks end by 20:30; networking until 21:00)
🎤 Format: In-person (onsite only)
🍽️ Food/Drinks: Provided by XPeng (pizza, non-alcoholic beverages)
📸 Media: Photos will be taken by both MunichNLP and XPeng. Talks may be recorded for YouTube (pending speaker approval).

​🎯 Why Attend?

• ​Learn from experts in generative audio, synthetic data, and multilingual LLMs.

• ​Connect with researchers, professionals, and enthusiasts in Munich’s AI/NLP community.

• ​Explore collaboration opportunities between academia and industry.