About Me

Here is Shifeng XIE.
I am a Master’s engineering student at Telecom Paris and the Polytechnic Institute of Paris. My research interests include large language models (LLMs), time‑series foundation models, graph neural networks and image reconstruction. I am familiar with network architectures such as Transformers, Channel‑wise attention mechanisms and UNet, and with training paradigms including contrastive, masked and joint learning.

I am planning to apply for PhD positions in 2026 and welcome opportunities worldwide!
If you are interested in my work, please feel free to reach out: shifeng.xie@telecom‑paris.fr.

Research Experiences

Time Series Foundation Models

Huawei Paris Noah’s Ark Lab, France
February 2025 – August 2025

Designed classification time‑series foundation models and demonstrated that high‑performance pretraining is possible using only synthetic data.
First author of “CauKer: Classification Time Series Foundation Models Can Be Pretrained on Synthetic Data Only,” accepted by ICML 2025 Workshop on Foundation Models for Structured Data (Best Time Series Paper).
Collaborated with Ievgen Redko and researchers from Huawei Paris.

In‑Context Learning and Mixture of Experts

Stellantis, France
July 2024 – January 2025

Investigated in‑context learning and gradient descent in transformers and mixture‑of‑experts (MoE) models.
First author of “The Initialization Determines Whether In‑Context Learning Is Gradient Descent,” published in Transactions on Machine Learning Research (TMLR), 2025.
Worked closely with Rui Yuan, Simone Rossi and Thomas Hannagan.

Graph Neural Networks and Graph Representation Learning

Télécom Paris (Polytechnic Institute of Paris), France
December 2023 – September 2024

Conducted research on variational graph contrastive learning and subgraph Gaussian embedding for self‑supervised graph representation.
First author of “Variational Graph Contrastive Learning,” accepted by NeurIPS 2024 Workshop on Self‑Supervised Learning – Theory and Practice.
First author of “Subgraph Gaussian Embedding Contrast for Self‑Supervised Graph Representation Learning,” accepted by ECML‑PKDD 2025.
Supervised by Jhony H. Giraldo.

Image Processing and High‑Dynamic‑Range Reconstruction

Xidian University, China
February 2023 – September 2023

Developed FTUnet for single HDR image reconstruction.
First author of “FTUnet: Feature Transferred U‑Net for Single HDR Image Reconstruction,” accepted by ACM Multimedia Asia (MMA) 2023, oral presentation.
Advised by Liu Yi.

Data Twin and Intelligent Healthcare

Xidian University, China
April 2021 – September 2021

Researched the feasibility of intelligent healthcare based on digital twin and data mining.
First author of “Feasibility Study of Intelligent Healthcare Based on Digital Twin and Data Mining,” accepted by CISAI 2021.

Education

Master in Engineering – Signal Processing for Artificial Intelligence

Télécom Paris & Polytechnic Institute of Paris, France
August 2023 – Present

Current grade: 15.3 / 20.

Summer Exchange Program – Machine Learning

McGill University, Canada
July 2021 – August 2021

Achieved grade: A.

Bachelor of Engineering – Electronic Information Engineering

Xidian University, China
August 2019 – May 2023

GPA: 3.8 / 4.0.
Rank: 4 / 97.

Selected Projects

Pretraining tiny vision & language models (≈ 2 billion parameters): Trained mixture‑of‑experts models using 8 A100 80 GB GPUs on the C4 and ImageNet datasets.
Fine‑tuning MoE language models with permutation symmetries and LoRA (patented): Applied permutation symmetries and Low‑Rank Adaptation to MoE models (e.g., Mistral, DeepSeek, Qwen) to improve efficiency.
Unsupervised face recognition with PCA and ICA: Implemented PCA and ICA to extract and modify facial features, enhancing recognition accuracy.
Self‑supervised learning for medical image classification: Developed contrastive learning methods on the MedMNIST database to learn representations without annotations.
Neural network parameter diffusion (patented): Compressed experts from MoE models into a latent space via autoencoders and trained latent diffusion models to generate new experts.

Honors and Awards

Best Time Series Paper, ICML 2025 Workshop on Foundation Models for Structured Data – for “CauKer: Classification Time Series Foundation Models Can Be Pretrained on Synthetic Data Only”.
Champion, Huawei France Tech Arena “Light Chaser” Competition (2024).
National Scholarship (China, 2021) – highest national‑level scholarship with a 0.2 % acceptance rate.
National Second Prize, Chinese Student Academic Physics Competition (2021).
Northwest Division Champion & First Prize, Shaanxi Province, Chinese Student Academic Physics Competition – best result ever at Xidian University.
Honorable Mention, International College Students Mathematical Contest in Modeling (2021).
First Provincial Prize, Mathematical Contest in Modeling for University Students (2021).

Skills & Service

Languages: Chinese (native), English (C1), French (B2).
Programming: Python (PyTorch, TensorFlow, JAX, SciPy, Pandas), JavaScript, Java, C++ and C.
Hardware: Arduino, STM32, SolidWorks, VHDL, ARM and RISC.
Professional service: Reviewer for NeurIPS 2024 Workshop on Compression, COLM 2025, and NeurIPS 2025.

News and Updates

August 2025: “CauKer: Classification Time Series Foundation Models Can Be Pretrained on Synthetic Data Only” won the Best Time Series Paper award at the ICML 2025 Workshop on Foundation Models for Structured Data.
July 2025: Published “The Initialization Determines Whether In‑Context Learning Is Gradient Descent” in TMLR.
December 2024: “Variational Graph Contrastive Learning” accepted at the NeurIPS 2024 Workshop on Self‑Supervised Learning – Theory and Practice.
May 2024: Champion in Huawei France Tech Arena “Light Chaser” Competition.
September 2023: Presented “FTUnet: Feature Transferred U‑Net for Single HDR Image Reconstruction” at ACM Multimedia Asia 2023 (oral).