Patrick Amadeus Irawan

I am a Ph.D. student at MBZUAI, advised by Alham Fikri Aji. Previously, I was a Research Engineer at Singapore Management University with Chong-Wah Ngo, and received my B.S. in Computer Science from Institut Teknologi Bandung, advised by Ayu Purwarianti.

I build evaluation and post-training methods for vision-language and video models, with a current focus on interactive environments. My overarching goal is to build robust, self-improving multimodal agents.

My work has appeared at CVPR, NAACL, EMNLP, COLING, and AACL, and received the Best Theme Paper Award at NAACL 2025.

I am looking for 2027 Research Scientist, Applied Scientist, and Research Engineer internships. Feel free to reach out!

[first.last]@mbzuai.ac.ae

Research Interests

• Multimodal reasoning and evaluation: finding where models fail, across tasks, languages, and cultures.
• Multimodal post-training: adding capabilities efficiently without forgetting existing ones.
• Interactive environments and world models: whether models can simulate goal-directed actions in games, sandboxes, and open worlds.

News

Selected Publications (Full Publications)

* equal contribution

Interactive Environments and World Models

• Ego2Act: Evaluating Goal-Directed Manipulation in Egocentric Video Generation
  Patrick Amadeus Irawan, Iskandar Muda Rizky Parlambang, Rava Maulana, Qinrong Cui, Erland Hilman Fuadi, ..., Shoubin Yu, Yue Zhang, Mohit Bansal, Alham Fikri Aji
  Preprint, 2026. [Project]

Multimodal Post-Training

• LinguDistill: Recovering Linguistic Ability in Vision Language Models via Selective Cross-Modal Distillation
  Patrick Amadeus Irawan, Erland Hilman Fuadi, Shanu Kumar, Alham Fikri Aji, Yova Kementchedjhieva
  Preprint, 2026. [arXiv]
• Towards Efficient and Robust VQA-NLE Data Generation with Large Vision-Language Models
  Patrick Amadeus Irawan, Genta Indra Winata, Samuel Cahyawijaya, Ayu Purwarianti
  International Conference on Computational Linguistics (COLING), 2025. Oral [arXiv] [Code]

Multimodal Reasoning and Evaluation

• Counting to Four is still a Chore for VLMs
  Duy Le Dinh Anh*, Patrick Amadeus Irawan*, Tuan Van Vo
  Computer Vision and Pattern Recognition Conference (CVPR) Workshop, 2026. [arXiv] [Code]
• Vision Language Models are Confused Tourists
  Patrick Amadeus Irawan*, Ikhlasul Akmal Hanif*, Muhammad Dehan Al Kautsar, Genta Indra Winata, Fajri Koto, Alham Fikri Aji
  Computer Vision and Pattern Recognition Conference (CVPR), 2026 Findings. [arXiv]
• M4-RAG: A Massive-Scale Multilingual Multi-Cultural Multimodal RAG
  David Anugraha, Patrick Amadeus Irawan, Anshul Singh, En-Shiun Annie Lee, Genta Indra Winata
  Computer Vision and Pattern Recognition Conference (CVPR), 2026. [arXiv]
• WorldCuisines: A Massive-Scale Benchmark for Multilingual and Multicultural Visual Question Answering on Global Cuisines
  Genta Indra Winata*, Frederikus Hudi*, Patrick Amadeus Irawan*, David Anugraha*, Rifki Afina Putri*, et al.
  North American Chapter of the Association for Computational Linguistics (NAACL), 2025. Best Theme Paper [arXiv] [Code]

Experience

• Ph.D. Student, MBZUAI, 2025 – present
• Research Engineer, Singapore Management University, 2025
• Software Engineer, IT Bauschmiede, 2024 – 2025
• Data Science and SWE Intern, Supertype, Blibli, Ruangguru, 2022 – 2023

Honors & Services

• Reviewer, CVPR 2026, ACL Rolling Review (ARR) 2025, NLPCC 2025
• Scholarship, Columbia Machine Learning Summer School, 2026
• Best Theme Paper Award, NAACL 2025
• OpenAI Micro Grant, 2025
• Winner, BCG SEA Emeralds Case Competition, 2024
• Finalist, Gemastik XV Data Mining Division, 2022