About Me

I am Shifeng Xie, a PhD student at Université Paris Cité working with Themis Palpanas. My research focuses on time series foundation models for classification, forecasting, and decision-making. I previously studied at Télécom Paris and the Institut Polytechnique de Paris.

I welcome PhD internship opportunities in 2027.

Contact:shifeng.xie@telecom-paris.fr.

Google Scholar · LinkedIn

I am currently building TSC-FM, a benchmark for time series classification foundation models.

Selected Publications

• CauKer: Classification Time Series Foundation Models Can Be Pretrained on Synthetic Data — ICLR 2026.
• Mantis: Lightweight Foundation Model for Time Series Classification — ICML 2026.
• Tabby: An Open Pretraining Recipe for Time Series Foundation Models — first-author preprint, 2026.
• Post-Training in Time Series Foundation Models: A Unifying Framework — first-author preprint, 2026.
• TSFMAudit: Data Contamination Auditing in Forecasting Time Series Foundation Models — equal-contribution first-author preprint, 2026.
• Rethinking Zero-Shot Time Series Classification: From Task-specific Classifiers to In-Context Inference — equal-contribution first-author preprint, 2026.

Full publication list

Research Experience

• Huawei Paris Noah’s Ark Lab (2025) — time series foundation models.
• Stellantis (2024–2025) — in-context learning and mixture of experts.
• Télécom Paris (2023–2024) — graph representation learning.