Zaber Ibn Abdul Hakim

114, 5th Ave

New York, NY

Hello! I’m Zaber, a Senior Data Scientist at Capital One. There I build evaluation frameworks for LLMs and agentic systems and improve the validation components of a production multi-agent system. I recently completed my M.S. in Computer Science (thesis) at Virginia Tech with a 4.00 GPA, advised by Dr. Chris Thomas. My research focuses on multimodal LLMs: multimodal information extraction, video and audio understanding, LLM post-training with reinforcement learning, and the security and robustness of multimodal models.

Before Capital One, I was an AI Research Intern at Comcast, where I built LLM-supervised clustering pipelines for large-scale conversation data. Before that, I was a Machine Learning Engineer at IQVIA under Srikanth Sankaran Iyer, where I fine-tuned and deployed the LLM behind IQVIA’s AI Assistant. Earlier, I was a research intern at Xu Lab, CMU.

My work has appeared at venues including ICLR, CVPR, ECCV, ACL, EMNLP and AAAI. Before research, I competed actively on Kaggle, DrivenData and other platforms. I earned the Kaggle Competition Master title with a peak rank of 385 out of 200,000+ competitors, and I won the IEEE Signal Processing Cup and the Kaggle Days World Championship 2022.

Interest

Information Extraction and Retrieval
Multi-Agent Systems and Evaluation
Large-Scale LLM Training

selected publications

EMNLP

M3EC: A Multimodal Multidocument Benchmark for Event Extraction and Event Coreference Resolution

Zaber Ibn Abdul Hakim, Najibul Haque Sarker, Aafiya Shamshad Hussain, and 5 more authors

In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing (EMNLP), 2026

HTML
ECCV

Reasoning-Guided Part-Level Visual Grounding via Reinforcement Learning

Kazi Sajeed Mehrab, Hani Alomari, Najibul Haque Sarker, and 4 more authors

In European Conference on Computer Vision (ECCV), 2026

arXiv
CVPR Highlight

Immunizing Models Against Harmful Long-Horizon Fine-Tuning via Contractive Optimization Dynamics

Najibul Haque Sarker, Zaber Ibn Abdul Hakim, Ali Asgarov, and 3 more authors

In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2026

HTML PDF
ICLR

SONICS: Synthetic Or Not–Identifying Counterfeit Songs

Md Awsafur Rahman^*, Zaber Ibn Abdul Hakim^*, Najibul Haque Sarker^*, and 2 more authors

International Conference on Learning Representations, 2025

Abs arXiv HTML

The recent surge in AI-generated songs presents exciting possibilities and challenges. While these tools democratize music creation, they also necessitate the ability to distinguish between human-composed and AI-generated songs for safeguarding artistic integrity and content curation. Existing research and datasets in fake song detection only focus on singing voice deepfake detection (SVDD), where the vocals are AI-generated but the instrumental music is sourced from real songs. However, this approach is inadequate for contemporary end-to-end AI-generated songs where all components (vocals, lyrics, music, and style) could be AI-generated. Additionally, existing datasets lack lyrics-music diversity, long-duration songs, and open fake songs. To address these gaps, we introduce SONICS, a novel dataset for end-to-end Synthetic Song Detection (SSD), comprising over 97k songs with over 49k synthetic songs from popular platforms like Suno and Udio. Furthermore, we highlight the importance of modeling long-range temporal dependencies in songs for effective authenticity detection, an aspect overlooked in existing methods. To capture these patterns, we propose a novel model, SpecTTTra, that is up to 3 times faster and 6 times more memory efficient compared to popular CNN and Transformer-based models while maintaining competitive performance. Finally, we offer both AI-based and Human evaluation benchmarks, addressing another deficiency in current research.
ICIP

Artifact: A large-scale dataset with artificial and factual images for generalizable and robust synthetic image detection

Md Awsafur Rahman^*, Bishmoy Paul^*, Najibul Haque Sarker^*, and 2 more authors

2023 IEEE International Conference on Image Processing (ICIP), 2023

Abs arXiv HTML Code

Synthetic image generation has opened up new opportunities but has also created threats in regard to privacy, authenticity, and security. Detecting fake images is of paramount importance to prevent illegal activities, and previous research has shown that generative models leave unique patterns in their synthetic images that can be exploited to detect them. However, the fundamental problem of generalization remains, as even state-of-the-art detectors encounter difficulty when facing generators never seen during training. To assess the generalizability and robustness of synthetic image detectors in the face of real-world impairments, this paper presents a large-scale dataset named ArtiFact, comprising diverse generators, object categories, and real-world challenges. Moreover, the proposed multi-class classification scheme, combined with a filter stride reduction strategy addresses social platform impairments and effectively detects synthetic images from both seen and unseen generators. The proposed solution outperforms other teams by 8.34% on Test 1, 1.26% on Test 2, and 15.08% on Test 3 in the IEEE VIP CUP at ICIP 2022.