Skip to main content


Office Location

440 Huntington Avenue
266 West Village H
Boston, MA 02115


Mahdi Khodayar is a PhD student in the Computer Science program Northeastern University’s College of Computer and Information Science, advised by Professor Ehsan Elhamifar. A native of Tehran, Iran, Mahdi earned both his bachelor’s degree and master’s degree from Khajeh Nasir Toosi University of Technology in Iran. Mahdi’s research areas include algorithms and theory, systems and networks, artificial intelligence, data science, machine learning, natural language processing, and information retrieval.

He is also a member of the Applied Machine Learning Group, Algorithms and Theory Group, and the Data Science and Analytics Laboratory. Prior to Northeastern, Mahdi worked on algorithm design and data structures. Mahdi’s current work includes machine learning and statistical pattern recognition.


  • MS, Khajeh Nasir Toosi University of Technology – Iran
  • BS, Khajeh Nasir Toosi University of Technology – Iran

About Me

  • Hometown: Tehran, Iran
  • Field of Study: Computer Science
  • PhD Advisor: Ehsan Elhamifar

What are the specifics of your graduate education (thus far)?

I study machine learning and statistical pattern recognition.

What are your research interests?

In my undergraduate program my studies were mainly focused on algorithm design and analysis and data structures. I was also interested in business process management systems and how we can compute the time and memory complexity of algorithms applied in this field. The graduate program was a great opportunity for me to study machine learning and pattern recognition, which helped me to work on learning deep architectures for artificial intelligence. My M.Sc. thesis was to build robust Deep Learning models for time series forecasting.

What’s one problem you’d like to solve with your research/work?

I am currently working on the incorporation of deep neural architectures with multitask learning for highly varying data.

What aspect of what you do is most interesting?

What mostly surprises me is the fact that capturing data from several measuring stations and using the spatial correlation can help multitask models perform significantly better than single task learners for time series predictors.

What are your research or career goals, going forward?

I love artificial Intelligence. I am involved with new problems which require probabilistic analysis techniques as well as machine learning models. I try to find effective solutions for such problems.