CV

PhD in Computer Science

• Focus on LLMs and machine translation.
• Supervised by Christof Monz.

Master of Science in Artificial Intelligence

• GPA: 4.0/4.0.

Bachelor of Science in Information System

Applied Scientist Research Intern

• Conducted research on test-time scaling for long-context, book-level machine translation, bridging research and production LLM systems.
• Designed and led a large-scale human evaluation pipeline with 1M annotated words and around $90K USD in annotation cost.
• Research output accepted to ACL 2026.

Research Intern

• Developed reasoning-based LLM evaluation models for translation quality estimation and refinement.
• Applied RLVR to improve evaluation reliability and alignment with human preferences using 7B-32B LLMs.
• Built a multi-role agentic LLM framework for evaluation and refinement pipelines.
• Research output accepted to ACL 2026.

What Does LLM Refinement Actually Improve? A Systematic Study on Document-Level Literary Translation

• ACL 2026
• Shaomu Tan, Dawei Zhu, Ke Tran, Michael Denkowski, Sony Trenous, Bill Byrne, Leonardo Ribeiro, Felix Hieber

Remedy-R: Generative Reasoning for Machine Translation Evaluation without Error Annotations

• ACL 2026
• Shaomu Tan, Ryosuke Mitani, Ritvik Choudhary, Qiyu Wu, Toshiyuki Sekiya, Christof Monz

ReMedy: Learning Machine Translation Evaluation from Human Preferences with Reward Modeling

• EMNLP 2025
• Shaomu Tan, Christof Monz

Neuron Specialization: Leveraging Intrinsic Task Modularity for Multilingual Machine Translation

• EMNLP 2024
• Shaomu Tan, Di Wu, Christof Monz

How Far Can 100 Samples Go? Unlocking Zero-Shot Translation with Tiny Multi-Parallel Data

• ACL Findings 2024
• Di Wu, Shaomu Tan, Yan Meng, David Stap, Christof Monz

Towards a Better Understanding of Variations in Zero-Shot Neural Machine Translation Performance

• EMNLP 2023
• Shaomu Tan, Christof Monz

Make Pre-trained Model Reversible: From Parameter to Memory Efficient Fine-Tuning

• NeurIPS 2023
• Baohao Liao, Shaomu Tan, Christof Monz

WSDM Cup - Multilingual Chatbot Arena

• Silver Medal - Ranked 19/950 teams
• Trained 7B-14B reward models for multilingual human preference prediction with DeepSpeed, vLLM, quantization, pseudo-labeling, and model merging.

Google American Sign Language Fingerspelling Recognition

• Gold Medal - Ranked 11/1,315 teams
• Designed and implemented a Conformer-Transformer architecture over sign-language landmarks with CTC and cross-entropy objectives.

WMT 2023 General Machine Translation Shared Task

• First Place on the English-Hebrew constrained track
• Built a compact encoder-decoder MT system that performed on par with GPT-4 5-shot prompting.

Interference and Knowledge Transfer in Multilingual Translation Models

• Co-presented with Prof. Christof Monz.

A Journey on Multilingual Neural Machine Translation

• Co-presented with Prof. Denis Paperno.

Program Committee / Reviewing

• Reviewed for ACL, EMNLP, EACL, NAACL, TASLP, WMT, and Multilingual Representation Learning.

Toward Explainable, Robust, and Actionable Translation Quality Estimation

Remedy-R: Large Reasoning Models for Machine Translation Evaluation

The Second Half of Machine Translation