Scholar
Adam Ibrahim
Google Scholar ID: rZvmqJsAAAAJ
Consultant & Mila, University of Montreal
Citations & Impact
Citations
693
H-index
9
i10-index
9
Publications
14
Co-authors
13
list available
Publications
Resume (English only)
Academic Achievements
- Published papers include 'Why Has Predicting Downstream Capabilities of Frontier AI Models with Scale Remained Elusive?' (ICML 2025), 'Zamba: A Compact 7B SSM Hybrid Model' (Technical report), and 'Simple and Scalable Strategies to Continually Pre-train Large Language Models' (TMLR). In the latter, Adam Ibrahim made equal contributions with other authors.
Research Experience
- Led the LLM team at H company and provided consulting services to Apple, Microsoft, AMD, Staples, Optina Diagnostics, and Bosch, supervising and mentoring research projects. Also consulted for Blackbox AI and Zyphra, notably co-leading the development of an efficient model based on a novel LLM architecture, Zamba-7B, in pretraining and finetuning foundation models.
Education
- PhD from Mila – Université de Montréal, supervised by Ioannis Mitliagkas and Irina Rish; Master's at the Four Eyes lab, University of California, Santa Barbara, focusing on human-computer interaction; Bachelor's in Joint Honours Mathematics and Physics at McGill University, researching theoretical physics (cosmology and string theory) under Robert Brandenberger and Keshav Dasgupta.
Background
- Research interests include multimodal machine learning, speech recognition, reinforcement learning, image generation, computer vision, large language models, and foundation models. During his PhD, he focused on addressing distributional shifts, which is crucial for efficient, fair, safe, and trustworthy machine learning algorithms.
Miscellany
- Enjoys all kinds of problem solving and is always happy to tackle problems in new areas of research.
