Home > People > Prashanth Mundkur

Prashanth Mundkur

Principal Researcher
Data Insight, Nokia Research Center, North America Lab

I am a Member of Research Staff at Nokia Research Center, North America Labs, which I joined in June 2010.  I work in the Data Insight team, building and programming distributed systems and infrastructure for large-scale data mining and machine learning.  One primary open-source product of this team is Disco, a MapReduce implementation built using Erlang, Python and OCaml.  I am interested in distributed systems, functional programming languages, and formal techniques.

Research Interests

Correctness in systems research often takes a back-seat to performance, leading to nasty problems that plague legacy systems (e.g. Linux) that are still in daily production use.  I am interested in approaches that take correctness in system design seriously, and in the tools (robust programming languages, domain-specific languages, model checkers and theorem provers) that one can employ in this task.

Personal Information

I have previously worked in WAN optimization and virtualization at Citrix Systems, and briefly in storage systems at Tintri.

Publications

  1. Disco: A Computing Platform for Large-scale Data Analytics,  Prashanth Mundkur, Ville Tuulos, Jared Flatow.  In Tenth ACM SIGPLAN Erlang Workshop, Tokyo, Japan, September 23, 2011.

Presentations

  1. Disco: using Erlang to implement Mapreduce, Prashanth Mundkur, Ville Tuulos, Jared Flatow. Presented at Commercial Users of Functional Programming 2011, Tokyo, Japan, September 24, 2011.