About me |
|
I am a Computer Vision researcher with the Intelligent Visual Analytics Lab (IVAL) in the Computer Vision Department at Mohamed bin Zayed University of Artificial Intelligence (MBZUAI), advised by Prof. Salman Khan. My research focuses on vision-language models (VLMs) for reasoning and OCR, with an emphasis on efficient and reliable real-world deployment. I’m pursuing an MSc in Computer Vision at MBZUAI, and I combine academic research with building production-facing ML systems. Previously, I was a Research Scientist Intern at Parameter Lab (in collaboration with NAVER LABS), an Applied Scientist at Monta AI, and a Software Engineer at And Africa Co., Ltd., with earlier ML internship experience at Scoville. Broadly, I build efficient, fast, and robust multimodal methods across OCR, VLMs, and on-device/mobile deployment. Email / Resume / Google Scholar / Github / Linkedin |
|
Selected Publications |
|
Dr.LLM: Dynamic Layer Routing in LLMs
Ahmed Heakl, Martin Gubri, Salman Khan, Sangdoo Yun, Seong Joon Oh Under Review Paper / Code Dr.LLM retrofits pretrained LLMs with lightweight per-layer routers that dynamically skip/execute/repeat layers under a compute budget, improving efficiency while preserving accuracy via supervised routing derived from MCTS. |
|
SVRPBench: A Realistic Benchmark for Stochastic Vehicle Routing Problem
Ahmed Heakl, Yahia Salaheldin Shaaban, Martin Takac, Salem Lahlou, Zangir Iklassov NeurIPS 2025 — Datasets and Benchmarks Paper / Code / Dataset SVRPBench is an open-source benchmark of 500+ urban-scale stochastic VRP instances featuring time-dependent congestion, probabilistic delays, and heterogeneous time windows, together with a full evaluation suite for learning-based solvers. |
|
GG: An LM Approach for CISC-to-RISC Transpilation with Testing Guarantees
Ahmed Heakl, Sarim Hashmi, Chaimaa Abi, Celine Lee, Abdulrahman Mahmoud EMNLP 2025 — Findings Paper / Code / Datasets and Models We introduce Guaranteed Guess (GG), a test-driven CISC-to-RISC transpiler using LLMs. GG achieves 99.4% functional accuracy on x86→ARMv8 translation and improves runtime and memory versus strong baselines through verification-centric generation and refinement. |
|
KITAB-Bench: Multi-Domain Arabic OCR and Document Understanding Bench
Ahmed Heakl*, Abdullah Sohail*, Mukul Ranjan*, Rania Hossam*, Ghazi Shazan Ahmad, Mohamed El-Geish, Omar Maher, Zhiqiang Shen, Fahad Khan, Salman Khan ACL 2025 — Findings Paper / Dataset / Code KITAB-Bench contains 8,809 Arabic pages spanning 9 domains and 36 sub-domains to evaluate OCR and document understanding. The benchmark highlights large gains from VLM-based OCR (≈60% lower character error than traditional OCR) while PDF-to-Markdown conversion peaks around ≈65%. |
|
LlamaV-o1: Rethinking Step-by-step Visual Reasoning in LLMs
Omkar Thawakar*, Dinura Dissanayake*, Ketan More*, Ritesh Thawkar*, Ahmed Heakl*, Noor Ahsan*, Yuhao Li, Mohammed Zumri, Jean Lahoud, Rao Muhammad Anwer, Hisham Cholakkal, Ivan Laptev, Mubarak Shah, Fahad Shahbaz Khan, Salman Khan ACL 2025 — Findings Paper / Code / Benchmark / Model We propose a curriculum framework and a benchmark with 4k+ reasoning steps across 8 categories for step-by-step visual reasoning. LlamaV-o1 improves over Llava-CoT by 3.8% while delivering ~5× faster inference. |
|
CASS: Nvidia to AMD Transpilation with Data, Models, and Benchmark
Ahmed Heakl, Sarim Hashmi, Gustavo Bertolo Stahl, Seung Hun Eddie Han, Salman Khan, Abdulrahman Mahmoud Under Review Paper / Code / Dataset CASS introduces the first large-scale dataset and benchmark (70k samples) for transpilation between Nvidia CUDA and AMD HIP/RDNA3 at both source and assembly levels. We train domain-specialized LLMs (up to 7B) that achieve strong functional preservation and substantially improve transpilation accuracy on challenging low-level code. |
|
VideoMolmo: Spatio-Temporal Grounding Meets Pointing
Ghazi Shazan Ahmad*, Ahmed Heakl*, Hanan Gani, Abdelrahman Shaker, Zhiqiang Shen, Fahad Shahbaz Khan, Salman Khan Preprint / Under Review Paper / Code / VPoS-Bench VideoMolmo is a spatio-temporal grounding framework that combines LLM-driven pointing with mask fusion (via SAM2). We introduce temporal attention and bidirectional mask propagation, curate a large video-caption dataset with dense object points, and propose VPoS-Bench to evaluate grounding, counting, and reasoning in real-world video domains. |
|
CAMEL-Bench: A Comprehensive Arabic LMM Benchmark
Sara Ghaboura*, Ahmed Heakl*, Omkar Thawakar, Ali Alharthi, Ines Riahi, Abduljalil Saif, Jorma Laaksonen, Fahad S. Khan, Salman Khan, Rao Muhammad Anwer NAACL 2025 Paper / Code CAMEL-Bench is a large-scale Arabic-centric LMM evaluation benchmark spanning diverse domains (OCR/Docs, medical, remote sensing, video, and more) with broad coverage for realistic Arabic multimodal reasoning. |
|
AIN: The Arabic INclusive Large Multimodal Model
Ahmed Heakl, Sara Ghaboura, Omkar Thawakar, Fahad Shahbaz Khan, Hisham Cholakkal, Rao Muhammad Anwer, Salman Khan Preprint Paper / Code AIN is a bilingual (Arabic/English) large multimodal model designed to close the Arabic LMM gap, targeting broad-domain multimodal understanding and strong Arabic performance across diverse tasks. |
|
How Good are Foundation Models in Step-by-Step Embodied Reasoning?
Dinura Dissanayake, Ahmed Heakl, Omkar Thawakar, Noor Ahsan, Ritesh Thawkar, Ketan More, Jean Lahoud, Rao Muhammad Anwer, Hisham Cholakkal, Ivan Laptev, Fahad Shahbaz Khan, Salman Khan Under Review Paper / Code Introduces the FoMER benchmark to evaluate step-by-step embodied reasoning for foundation multimodal models, analyzing action validity and reasoning correctness across diverse tasks and embodiments. |
News |
Academic Service |