Kamal Hakimzadeh Harirbaf

I am a Ph.D. student in the group of distributed systems, SCS department, School of  Electrical Engineering and Computer Science (EECS) of KTH.

My broad research interests are Distributed Systems, Cloud Computing, Elasticity, and Machine Learning. More specifically, I work on the related problems in Auto-Scaling of large-scale systems in the Cloud.

Currently, I am doing an internship at Bell Labs in Dublin/Ireland. During my internship, I am working on a novel solution for application agnostic auto-scaling using Reinforcement Learning. In my former project, I worked on a novel functional model for configuration management (CM) as we auto-scale systems.

Research Projects

I am the founder of the Karamel, an open-source project for cloud-based provisioning of large-scale distributed systems. In addition to the engine, we have developed other sub-modules for Karamel that are open-sourced too. HoneyTap is Karamel's auto-scaling system, TableSpoon is a distributed monitoring system for resource utilization,  Kandy is a system for provisioning of transient servers for maximizing the reliability. 

Earlier in my master thesis, I had worked on a project for scaling out the Hadoop Distributed File Systems (HDFS). My contribution to that work is the hierarchical and pessimistic row level concurrency control for maintaining the metadata consistency (source code, publication). This work is continued by other students and it is called HopsFS for improving the performance. (source code, publication)

Teaching

Since the start of my Ph.D., I have assisted in designing homework, labs, tutorials and also helping students for the following courses in our department. 

ID1020 - Algorithms and Data Structures

ID2223 - Scalable Machine Learning and Deep Learning

ID1217 - Concurrent Programming