Written by: Matthew Busekroos | Produced by: Nate Caldwell

Originally born in Iran, PhD candidate Mehrdad Khani studied at Sharif University of Technology double majoring in computer science and electrical engineering, prior to CSAIL and MIT. Khani said he applied to MIT's Computer Science and Artificial Intelligence Laboratory to continue his education because he wanted to learn from the best mentors, collaborators, and researchers in the world. He added the opportunity to be part of a community of leading researchers who push boundaries and make significant contributions in the field was highly appealing to him.

Khani works in the Systems Group in the lab alongside his advisor Professor Mohammad Alizadeh. His primary research interests are in the areas of computer networks, applied machine learning, and computer systems, with a focus on networked machine learning systems.
“Working with Professor Mohammad Alizadeh and my research group has been a rewarding experience,” Khani said. 
Khani added that the research group has a friendly, collaborative atmosphere and Professor Alizadeh is a great mentor and researcher. 
“He is knowledgeable, experienced, and always willing to share his insights and guidance with the group,” Khani said. “One of the most important lessons I have learned from him is the value of staying curious and open-minded in research. Mohammad is supportive of rethinking established practices and exploring new ideas, which has enabled me to work on several challenging innovative research proposals.”
Khani’s primary research interest is in the area of networked machine learning systems. One of Khani’s current projects focuses on using machine learning algorithms to improve the quality of video conferencing. With the growing trend of remote work and virtual communication, Khani said there is a high demand for clear, stable, and reliable video conferencing systems. 
However, traditional systems can be limited by factors such as bandwidth constraints and network latency. Khani said this project aims to address such challenges by developing new algorithms for video compression, denoising, and super-resolution that can be run in real-time on edge devices such as laptops and smartphones. Additionally, he said the group is working on techniques to enhance the stability and reliability of video conferencing in the face of network fluctuations.
“Networked machine learning systems have the potential to transform various industries by allowing for large-scale, real-time, and distributed processing of data,” he said. “These systems have numerous benefits, including scalability, improved accuracy, increased efficiency, and collaboration. Networked ML systems can handle massive amounts of data, provide up-to-date insights, and achieve a higher level of accuracy compared to single-machine systems. By reducing the time and resources required for data processing, networked ML systems make it more cost-effective and efficient. Furthermore, they enable collaboration and information sharing between organizations, leading to more effective decision-making and problem-solving.”
Khani said his motivation is to bring machine learning systems from the lab to the real world. 
“I believe that by addressing the computer systems challenges associated with machine learning, we can make a profound impact on our daily lives,” Khani said. “From healthcare to transportation, the potential benefits of machine learning are vast. I am passionate about developing scalable, efficient, and secure systems that process and analyze large amounts of data, bringing the benefits of machine learning to more people and organizations.”

Khani said his dream job is one where he can continue to create real-world applications of machine learning. He said he is passionate about working on projects that have the potential to make a real difference, whether in healthcare, education, transportation, or another field. 

“My goal is to create accessible and scalable machine learning systems that can be used to solve real-world problems and improve our world in meaningful ways,” he said. 
Khani regularly shares research updates, including papers, codes, and other relevant materials on his personal website:  http://people.csail.mit.edu/khani/