I am a Research Staff Member at IBM T.J. Watson Research Center. My research interests focus on designing algorithms and platforms to efficiently manage emerging heterogeneous workloads on cloud infrastructures. Currently I am interested in applying market-based concepts to maximize the resource utilization and prioritize smartly between workloads (e.g., data analytics, deep learning) on container clouds.
Before joining IBM Research I was a postdoc at Chalmers University of Technology, Sweden, where I analyzed the performance of cloud stream processing frameworks (Spark and Flink) for IoT use cases, in particular intelligent vehicular systems. Prior to this position I was a postdoc at VU Amsterdam, The Netherlands, working on aspects of efficiently supporting diverse applications (e.g. in-memory data stores, HPC applications) in the cloud at minimum costs for users and provider.
I obtained my PhD from University of Rennes 1, France, in 2013. During my PhD studies I was a Research Engineer at INRIA Rennes and EDF R&D. In my PhD work I used market mechanisms to arbitrate the resource access on private clouds between HPC applications with different resource requirements while considering that users might have different performance objectives and importance for executing them. I finished my BS (2008) and MS (2010) at the University Politehnica of Bucharest, Romania. During my MS studies I proposed and implemented a self-healing active replication solution for distributed grid services by using atomic broadcast over peer to peer overlays. I also had the chance to briefly work in the Molecular Dynamics group from University of Groningen on large scale parallelization of MD simulation algorithms on supercomputers.