Hello, I'm

Xianda Du

Computer Engineering & AI Researcher

University of Waterloo · 4.0/4.0 GPA
Researching NLP, Vision-Language Models & Information Retrieval

Xianda Du

About Me

I am an undergraduate student in Computer Engineering with an Option in Artificial Intelligence at the University of Waterloo, maintaining a perfect 4.0 GPA. My research interests lie at the intersection of Natural Language Processing, Vision-Language Models, and Information Retrieval.

I have had the privilege of working with Prof. Renee Sieber (McGill University), Prof. Wenhu Chen (University of Waterloo), and Prof. En-Hui Yang (University of Waterloo) on projects spanning benchmark construction, retrieval-augmented generation, and adversarial robustness.

I am actively seeking Master's and research positions where I can deepen my contributions to NLP and multimodal AI research.

4.0 GPA (out of 4.0)
3 Research Labs
ACL 2025 Publication
SIGIR 2026 Reviewer

Research Experience

Research Assistant

McGill University · Advisor: Prof. Renee Sieber

Sep 2024 – Present
  • WXImpactBench (ACL 2025 Findings): Built a benchmark for evaluating LLMs on multi-label classification and retrieval of societal impacts from historical weather disasters; proposed a unified evaluation pipeline.
  • WeatherArchive-Bench: Constructed a 1M+ segment benchmark for retrieval and climate vulnerability assessment aligned with IPCC frameworks.
  • WXChat: Developed a large-scale archival newspaper dataset via adaptive preprocessing, layout-aware segmentation, and LLM-based post-correction.

Research Assistant

University of Waterloo · Advisor: Prof. Wenhu Chen

Oct 2025 – Present
  • VIEScore2: Developing a unified supervised fine-tuning framework for vision-language models on image quality evaluation with visual grounding.

Undergraduate Research Assistant

University of Waterloo · Advisor: Prof. En-Hui Yang

Jan 2025 – Apr 2025
  • Noise Injection Analysis: Analyzed effects of Gaussian noise and adversarial perturbations (FGSM) on intermediate representations in pretrained ResNet-18.

Publications & Service

ACL 2025 Findings

WXImpactBench: A Benchmark for Evaluating LLMs on Historical Disruptive Weather Impacts

Xianda Du*, et al. (*Co-first author)

The first benchmark combining NLP and meteorology for LLMs' understanding of historical disruptive weather impacts. Benchmarked 12 LLMs on multi-label classification and ranking-based QA tasks across 1.7K annotated samples.

In Preparation (ICLR-targeted)

Climate-Domain RAG Benchmark

A benchmark to assess retriever accuracy, climate adaptation, and generation quality for improving LLM reliability in domain-specific tasks.

Academic Service

SIGIR 2026 Resource Track — Reviewer

Professional Experience

Research Assistant (NLP Benchmark)

McGill University – RBC Borealis AI

Oct 2024 – Oct 2025
  • Co-first authored WXImpactBench (ACL 2025), combining NLP and meteorology for understanding historical weather impacts.
  • Processed 53K+ OCR-scanned articles with GPT-4o post-OCR correction, LDA topic modeling, and expert curation.
  • Benchmarked 12 LLMs with multi-label classification and ranking-based QA evaluation using a sliding window re-ranking pipeline.

RAG Project Technical Lead

University of Waterloo, WatAI

Apr 2025 – Dec 2025
  • Built a production RAG flow with LangChain supporting multiple LLM backends, hybrid retrieval (dense + BM25), re-ranking, and grounded citations.
  • Implemented agent tools, model routing by speed/cost/capability, and evaluation with Recall@k, nDCG, and LLM-as-judge faithfulness.

Platform Developer

University of Waterloo, ECE Department

Sep 2024 – Dec 2024
  • Fine-tuned Hugging Face sentiment analysis models using TensorFlow/PyTorch, achieving 80% accuracy on media content classification.
  • Built a pipeline to convert embeddings from PostgreSQL on GKE into a neural network for article classification.

Technical Skills

Machine Learning & AI

PyTorchTensorFlowKerasTransformers Scikit-learnRAGCNNLSTM XGBoostLightGBMHPO

Programming Languages

PythonJavaScript/TypeScriptC/C++ JavaSQLBashRMATLAB

Frameworks & Tools

ReactFastAPILangChainDocker KubernetesGCPAWSGit SupabaseQdrant

Data & Analysis

NumPyPandasOpenCVMatplotlib HuggingFaceJupyter

Languages

Chinese (Native)English (Fluent) Japanese (Intermediate)French (Beginner)

Get in Touch

I am actively looking for Master's and research positions in NLP, multimodal AI, and information retrieval. Feel free to reach out.