Zuhair Hasan

Zuhair Hasan Shaik

I am a researcher at Singapore Management University (SMU) SCIS, working with Prof. Sun Jun. My research focuses on AI Safety, Post-training & Alignment, and Multimodal Systems. I graduated from IIIT Dharwad in 2025, where I worked with Prof. Sunil Saumya and Prof. Deepak K. T.. Previously, I was a researcher at MBZUAI, Abu Dhabi with Prof. Yova Kementchedjhieva, and a research intern at Microsoft Research (M365), Bengaluru with Dr.Minghua Ma. I have also worked with the FLaMe.nlp Research Lab and LCS2 Lab with Prof. Md. Shad Akhtar.

To know more, refer to my resume or drop me an email!

Sept 2026: I've joined Singapore Management University as a research Engineer and will be working with Prof. Sun Jun.
Sept 2025: Received the NeurIPS 2025 Scholar Award.
Sept 2025: Our paper Redefining Experts: Interpretable Decomposition of Language Models for Toxicity Mitigation got accepted at NeurIPS 2025.

Aug 2025: Our paper M-Help on Mental Health Help-Seeking Behavior Detection got accepted at EMNLP 2025 Findings.
Aug 2025: I've Joined MBZUAI as a Research Engineer after completing my graduation from IIIT Dharwad and concluding my tenure at Microsoft Research India.
Jan 2025: Our paper on multimodal detection of mental health disorders in social media has been accepted at TheWebConf (WWW) 2025.
Jan 2025: I've been invited to Google DeepMind Research Symposium 2025 aka Google Research week.
Dec 2024: I've joined Microsoft Research India as a research intern will be working with M365 team.
Sept 2024: Our paper LaRA on Cross modal LLM with speech got accepted at EMNLP 2024 Findings.
Sept 2024: Started my research internship at FLaMe.nlp Research Lab.
June 2024: Our paper on Emotion recognition in multi-party conversations (code-mixed) and emotion trigger dynamics, got accepted at SemEval 2024 co-held with NAACL '24.
May 2024: Started my research internship at LCS2 Research Lab.
March 2024: Our paper on Hate Speech Detection in Telugu-English Code-Mixed Text got accepted at DravidianLangTech 2024 co-held with EACL '24.
Feb 2024: Our team, FeedForward, secured 3rd place in Tasks 1 and 2, and 5th place in Task 3 of the SemEval 2024 EDIREF competition.
Dec 2021: Started my Bachelor degree (BTech) at IIIT Dharwad, India.

Structure and Psycho-social Safety as Language Models Move Closer to Human
DS4MH| Data Science for Mental Health (DS4MH) @ The Alan Turing Institute
slides

Redefining Experts: Interpretable Decomposition of Language Models for Toxicity Mitigation
Zuhair Hasan, Abdullah Mazhar, Aseem Srivastava, Md Shad Akhtar
NeurIPS '25 | Conference (A*) | Advances in Neural Information Processing Signals 2025
paper code slides

M-HELP: Using Social Media Data to Detect Mental Health Help-Seeking Signals
MSVPJ Sathvik*, Zuhair Hasan*, Vivek Gupta
EMNLP (Findings) '25 | Conference (A*) | Conference on Empirical Methods in Natural Language Processing 2025
paper dataset

Figurative-cum-Commonsense Knowledge Infusion for Multimodal Mental Health Meme Classification
Abdullah Mazhar, Zuhair Hasan, Aseem Srivastava, Polly Ruhnke, Lavanya Vaddavalli, Sri Keshav Katragadda, Shweta Yadav, Md Shad Akhtar
WWW '25 | Conference (A*) | International World Wide Web Conference 2025
paper code slides

LaRA: Large Rank Adaptation for Speech and Text Cross-Modal Learning in Large Language Models
Zuhair Hasan, Pradyoth Hegde, Prashanth Bannulmath, Deepak KT
EMNLP (Findings) '24 | Conference (A*) | Conference on Empirical Methods in Natural Language Processing 2024
paper code slides