Current Ongoings
I'm currently looking for a full-time research internship starting January 2026, I'm also open for remote research collaborations this fall, starting on September 2025.
2025 Highlights
Went from ML fundamentals to engineering LLM training and data pipelines from scratch.
Contributed to Large Multimodal Model research as an intern at QCRI.
Launched independent research into tokenizer-induced data leakage.
ABOUT ME
Dedicated, ambitious, always an amateur.
"In order for connection to happen, we have to allow ourselves to be seen— really seen."I'm Fadi Benzaima. I build the systems that power and explain large-scale AI. My work sits at the intersection of engineering and research.
I'm driven by a need to move beyond alchemy and toward a principled understanding of how these models work. Whether it's by engineering a custom GPU kernel or designing a controlled experiment, my goal is the same: the quest for good explanations.This website is my open-source notebook. Here, I document my process, share insights from my projects, and explore the questions that drive my research.
Journey timeline
This is a small breakdown of my journey so far!
Deep Dive into ML
Transitioned my focus to AI, building a strong theoretical foundation with Berkeley's advanced ML curriculum.
The Builder Phase
Built and open-sourced full-stack LLM infrastructure, from data curation pipelines to multi-GPU training loops.
Research Internship @ QCRI
Joined the Qatar Computing Research Institute to contribute to state-of-the-art research in Large Multimodal Models.
Systems & Specialization
Focused on low-level systems, optimizing GPU performance with Triton and pursuing independent research.