About me

Hi, I’m Arash Asgari, a machine learning researcher and engineer based in Toronto with over three years of experience spanning AI research, data pipelines, and deployed ML systems. I work on evaluating and improving large language and vision-language models, with a focus on fairness, robustness, and efficient inference. I completed my MSc in Computer Science at York University in August 2026, where I worked in the Responsible AI Lab under Dr. Laleh Seyyed-Kalantari.

My experience connects research with implementation: I developed LLM evaluation pipelines for my first-author ACL 2026 paper, co-developed ChartBias to evaluate bias across 12 vision-language models, and helped curate DistilKaggle, a dataset of over 542,000 computational notebooks. At the Vector Institute, I evaluated LLM knowledge-editing methods and optimized inference on compute clusters. As a Machine Learning Engineer at Drivee, I built cloud image-processing pipelines, optimized models with ONNX and quantization, and developed conversational AI agents.

I use Python, SQL, PyTorch, and Hugging Face to build reproducible experiments and ML workflows, with additional experience in Spark, cloud platforms, and data analysis. My background combines model development, statistical evaluation, dataset curation, and software engineering. Before York, I studied Computer Engineering at Sharif University of Technology and K.N. Toosi University of Technology.

I’m exploring opportunities as a Research Engineer, Research Scientist, Machine Learning Engineer, Data Scientist, or Data Engineer, where I can turn research ideas and complex data into reliable, useful systems.

Feel free to email me at arash.asgari.m@gmail.com or check my CV.

📣 Recent News