About

I am a Master's student in Computer Science at the University of Macau, working under the supervision of Prof. Derek F. Wong at the NLP & Portuguese–Chinese Machine Translation Laboratory (NLP2CT). My research focuses on explainable and trustworthy artificial intelligence for large language models, with particular emphasis on hallucination detection and the evaluation of LLM-generated text. I am also interested in multilingual and low-resource machine translation, including cross-lingual benchmarking and evaluation methodologies. I am fortunate to be closely mentored by Junchao Wu, whose guidance has been instrumental in shaping my research journey.
Explainable AI Trustworthy AI LLM Hallucination Detection LLM-Generated Text Evaluation Multilingual NLP Low-Resource MT Cross-Lingual Benchmarking Efficient NLP Systems

News

2026 Submitted IndicDetect to ACL Rolling Review (ARR) — manuscript under review.
Sep 2025 Started MSc in Computer Science at the University of Macau. Also began as Teaching Assistant for Discrete Structures and Software Project Management.
May 2025 Completed B.Tech (Honors) in CSE from KL University — GPA 9.00/10.00.
Mar 2024 Joined the NLP2CT Laboratory as a Research Assistant.

Publications

C = Conference · J = Journal · S = In Submission · T = Thesis

  1. S.1
    IndicDetect: Evaluating Cross-Lingual LLM-Generated Text Detection for Hindi, Telugu, and Tamil
    Bhaskar Ganesh Devalla,Junchao Wu, Nilesh Dokuparthi, Greeshma Yaluru, Tatiana Muniz Rodriguez, Lidia S. Chao, and Derek F. Wong. (2026).
    Manuscript submitted to ACL Rolling Review (ARR)

Research & Teaching Experience

NLP & Portuguese–Chinese Machine Translation Lab (NLP2CT)
Research Assistant — University of Macau
Macau SAR, China
  • Implemented and evaluated Transformer-based models (BERT, RoBERTa, GPT-style) for multilingual machine translation; developed test plans, triaged model failures, and iterated on configurations across low-resource language pairs.
  • Built automated evaluation pipelines (BLEU, ROUGE, WER, BERTScore) into reproducible, containerized workflows using Git and Docker, with experiment tracking via MLflow and Weights & Biases.
  • Conceived and led IndicDetect, a cross-lingual benchmark for LLM-generated text detection across Hindi, Telugu, and Tamil; drove dataset construction, evaluation framework design, and analysis of zero-shot and fine-tuned detector performance.
  • Co-authored a manuscript submitted to ACL Rolling Review (ARR) 2026, contributing experimental design, writing, and result analysis.
University of Macau
Teaching Assistant
Macau SAR, China
  • Conducted weekly tutorial sessions for undergraduate classes in Discrete Structures and Software Project Management, reinforcing algorithmic problem-solving and systematic software quality practices.
  • Supported assessment design and grading workflows, applying test coverage principles and edge-case analysis to evaluate student submissions effectively.
  • Prepared supplementary instructional materials and coordinated grading activities with the course instructor.

Education

University of Macau
Master of Science in Computer Science
Macau SAR, China
GPA: 4.00 / 4.00
Supervisor: Prof. Derek F. Wong
Koneru Lakshmaiah Education Foundation (KL University)
Bachelor of Technology in Computer Science and Engineering (Honors)
Guntur, India
GPA: 9.00 / 10.00

Selected Projects

IndicDetect: Cross-Lingual LLM-Generated Text Detection Benchmark Under Review · ACL ARR 2026
  • Constructed a multilingual dataset spanning human- and LLM-generated text in Hindi, Telugu, and Tamil; designed annotation guidelines and inter-annotator agreement protocols to ensure label quality.
  • Evaluated multiple detection models, including fine-tuned BERT variants and zero-shot GPT-based detectors, across languages; identified cross-lingual performance gaps and proposed transfer learning strategies to address them.
  • Packaged the benchmark with a reproducible evaluation harness enabling one-command re-evaluation.
Python · Hugging Face Transformers · PyTorch · scikit-learn · pandas
Code kept private while paper is under review
Emotion Aid: Emotion Speech Recognition for Disordered Speech
  • Developed a deep learning pipeline for emotion recognition from disordered speech; extracted MFCC, prosodic, and spectral features using Librosa and trained CNN-LSTM classifiers on clinical audio datasets with TensorFlow.
  • Automated the end-to-end evaluation pipeline covering feature extraction, model inference, and multi-class performance reporting, improving experiment reproducibility and reducing manual effort.
  • Conducted systematic failure analysis across emotion categories and applied targeted data augmentation strategies, including pitch shifting, time stretching, and noise injection, to address class imbalance.
Python · Flask · TensorFlow · Scikit-learn · Librosa · MySQL
View on GitHub

Technical Skills

Programming Python, C/C++, Java, Bash DL & LLM Frameworks PyTorch, TensorFlow/Keras, Hugging Face Transformers, Fairseq, OpenNMT; Transformer architectures (BERT, RoBERTa, GPT-style); BPE, SentencePiece NLP & Speech SpeechBrain, ESPnet, Librosa, NLTK, SpaCy; MT evaluation: BLEU, WER, ROUGE, BERTScore; cross-lingual benchmarking, low-resource NLP MLOps & Repro. MLflow, Weights & Biases, Docker, Git, CI pipeline integration; pytest, shell-based test automation Data & Analysis NumPy, Pandas, SciPy, Matplotlib, Seaborn Tools Linux/HPC cluster environments, LaTeX, Jupyter, VS Code Languages English (Advanced), Hindi (Fluent), Telugu (Native), Mandarin (Conversational)

Academic Service

Student Reviewer (under faculty supervision) — IEEE/ACM Transactions on Audio, Speech, and Language Processing

References

Derek F. Wong
Professor and Associate Head
Department of Computer and Information Science
Faculty of Science and Technology, University of Macau
Loading...
Suryakanth Veerashetty Gangashetty
Professor
Department of Computer Science and Engineering
Koneru Lakshmaiah Education Foundation (KL University)
Loading...
Flag Counter