LLM Evaluation & Alignment大语言模型评测与对齐
Reliable evaluation protocols, safety alignment, and machine-generated text detection for multilingual large language models.面向多语言大语言模型的可靠评测协议、安全对齐与机器生成文本检测。
MSc Researcher · NLP²CT Lab硕士研究生 · NLP²CT 实验室
Towards trustworthy multilingual large language models.迈向可信赖的多语言大语言模型
I am a Master's student in Computer Science at the University of Macau, working in the NLP²CT Laboratory under the supervision of Prof. Derek F. Wong. My research focuses on trustworthy multilingual large language models, particularly LLM evaluation and alignment, internal representation analysis, and robustness in cultural and low-resource settings. I lead the IndicDetect benchmark and the CultMetric evaluation framework, and I am closely mentored by Junchao Wu. I am always open to collaborating on trustworthy multilingual LLMs and low-resource NLP, so please feel free to reach out by email.
我是澳门大学计算机科学专业的硕士研究生，在自然语言处理与葡中机器翻译实验室（NLP²CT）由 Derek F. Wong 教授指导开展研究。我的研究聚焦于可信赖多语言大语言模型，尤其关注大语言模型评测与对齐、内部表征分析，以及文化与低资源场景下的鲁棒性。我主导 IndicDetect 基准与 CultMetric 评测框架，并得到吴俊超师兄的悉心指导。我始终乐于在可信赖多语言大语言模型与低资源自然语言处理方向开展合作，欢迎通过邮件与我联系。
01
Research Interests研究兴趣
My primary research focus is Trustworthy Multilingual LLMs - developing and evaluating reliable language models across diverse cultural and low-resource settings. 我的主要研究方向是可信赖多语言大语言模型--面向多元文化和低资源场景开发与评测可靠的语言模型。
Internal Representation Analysis内部表征分析
Studying how language models encode multilingual knowledge and how their internal representations affect generation, reasoning, and evaluation.研究语言模型如何编码多语言知识，以及内部表征如何影响生成、推理与评测。
Robustness & Low-Resource NLP鲁棒性与低资源自然语言处理
Robustness under distribution shift, cultural fidelity, and evaluation for low-resource languages including Hindi, Telugu, and Tamil.分布偏移下的鲁棒性、文化忠实度，以及面向印地语、泰卢固语和泰米尔语等低资源语言的评测。
02
News动态
IndicDetect has been accepted to the Findings of EMNLP 2026.IndicDetect 已被 EMNLP 2026 Findings 接收。 New最新
Submitted CultMetric, a gloss-grounded cultural-fidelity evaluation framework for machine translation, to ACL Rolling Review (ARR).向 ACL Rolling Review (ARR) 投稿 CultMetric：一套基于词条释义的机器翻译文化忠实度评测框架。
Submitted IndicDetect to ACL Rolling Review (ARR).向 ACL Rolling Review (ARR) 投稿 IndicDetect。
Started MSc in Computer Science at the University of Macau, appointed Teaching Assistant for Discrete Structures and Software Project Management.开始在澳门大学攻读计算机科学硕士学位，担任《离散结构》与《软件项目管理》课程助教。
Graduated with B.Tech (Honors) in Computer Science and Engineering from KL University.获 KL 大学计算机科学与工程荣誉工学学士学位。
Joined the NLP²CT Laboratory as an undergraduate exchange researcher.加入 NLP²CT 实验室，担任本科交流研究员。
03
Publications论文发表
LLM Generation & Detection大语言模型生成与检测
04
Experience经历
-
Research Assistant研究助理
May 2024 - Present2024年5月 - 至今NLP²CT Laboratory · University of MacauNLP²CT 实验室 · 澳门大学
- Lead IndicDetect, a cross-lingual benchmark for LLM-generated text detection across Hindi, Telugu, and Tamil.
- Built CultMetric, a reference-free framework for evaluating cultural fidelity in machine translation.
- Develop reproducible multilingual evaluation pipelines using BLEU, ROUGE, BERTScore, Docker, MLflow, and Weights & Biases.
- Contributed experimental design, writing, and analysis to manuscripts submitted to ACL Rolling Review (ARR) 2026.
- 主导 IndicDetect：面向印地语、泰卢固语与泰米尔语的跨语言大语言模型生成文本检测基准。
- 构建 CultMetric：一套无参考译文的机器翻译文化忠实度评测框架。
- 使用 BLEU、ROUGE、BERTScore、Docker、MLflow 与 Weights & Biases 开发可复现的多语言评测流程。
- 参与 ACL Rolling Review (ARR) 2026 投稿论文的实验设计、写作与结果分析。
-
Teaching Assistant助教
Sep 2025 - Jun 20262025年9月 - 2026年6月- Facilitated weekly tutorials for Discrete Structures , Software Project Management and Applied Natural Language Processing.
- Designed assessment rubrics and worked with faculty on instructional materials and grading for large computer-science classes.
- 为本科课程《离散结构》与《软件项目管理》主持每周辅导课。
- 设计评分标准，并与教师协作完成教学材料设计和大规模课程评分工作。
05
Education教育背景
Aug 2025 - Present2025年8月 - 至今
Master of Science in Computer Science计算机科学理学硕士
University of Macau澳门大学
Macau SAR, China中国澳门
Supervisor导师Prof. Derek F. WongDerek F. Wong 教授
Aug 2021 - May 20252021年8月 - 2025年5月
Bachelor of Technology (Honors) in Computer Science and Engineering计算机科学与工程荣誉工学学士
Koneru Lakshmaiah Education Foundation (KL University)Koneru Lakshmaiah 教育基金会（KL 大学）
Guntur, India印度 贡土尔
06
Professional Skills专业技能
Programming Languages编程语言
- Python
- C/C++
- Java
- SQL
- R
- Bash
Evaluation Metrics评测指标
- AUROC
- F1
- BLEU
- ROUGE
- BERTScore
- WER
Deep Learning & LLM Frameworks深度学习与大语言模型框架
- PyTorch
- TensorFlow / Keras
- Hugging Face Transformers
- QLoRA
- Fairseq
- OpenNMT
- Transformer architecturesTransformer 架构
- BERT
- RoBERTa
- XLM-RoBERTa
- GPT-styleGPT 系列
- BPE
- SentencePiece
NLP & Speech自然语言处理与语音
- NLTK
- SpaCy
- SpeechBrain
- ESPnet
- Librosa
- MFCC, prosodic & spectral featuresMFCC、韵律及频谱特征提取
- CNN-LSTM classifiersCNN-LSTM 分类器
- Cross-lingual benchmarking跨语言基准测试
- Low-resource NLP低资源自然语言处理
MLOps & ReproducibilityMLOps 与可复现性
- MLflow
- Weights & Biases
- Docker
- Git
- CI pipeline integrationCI 流水线集成
- pytest
- Shell-based test automation基于 shell 的测试自动化
Data & Analysis数据与分析
- pandas
- NumPy
- SciPy
- scikit-learn
- Matplotlib
- Seaborn
Developer Tools开发工具
- Linux / HPC clustersLinux / 高性能计算集群
- Google Cloud Platform
- LaTeX
- Jupyter
- VS Code
Languages语言
- English英语 Fluent流利
- Telugu泰卢固语 Native母语
- Hindi印地语 Fluent流利
- Mandarin普通话 Conversational日常交流
07
Academic Service学术服务
Student Reviewer学生审稿人
2023 - 2024IEEE/ACM Transactions on Audio, Speech, and Language Processing (TASLP)
Reviewed three manuscripts under the supervision of Prof. Suryakanth V. Gangashetty.在 Suryakanth V. Gangashetty 教授指导下审阅了三篇论文。
Get in touch联系我
Let's collaborate.欢迎合作交流
I am always open to collaborating on trustworthy multilingual LLMs and low-resource NLP. Feel free to reach out by email.我始终乐于在可信赖多语言大语言模型与低资源自然语言处理方向开展合作，欢迎通过邮件与我联系。