PhD Candidate
University of Toronto & Vector Institute
(✦ I publish under my legal name: Ahmadreza)
If you’re an undergraduate or master’s student in Toronto with a strong foundation in VLMs/LLMs and PyTorch, and you’re interested in my research, feel free to reach out for collaboration.
I do research on vision–language models: training them to reason more reliably (especially via RL/post-training with verifiable, self-supervised signals) and making inference more efficient (elastic-depth/looped transformer ideas)