Melanie Kambadur

Software Engineer at Oscar Health Insurance

Ph.D., Computer Science

firstnamelastname@gmail.com
New York, NY

About

I recently completed my Ph.D. at Columbia University's Computer Science Department and now work at Oscar, a health insurance startup in New York City.

During my Ph.D., my research focused on measurement and performance analysis at the intersection of computer software and hardware, in particular as related to energy efficiency, datacenter scheduling, parallel computing and multiprocessors, compilers, GPUs, computer vision, machine learning, and memory architectures. I worked with Professor Martha Kim in the ARCADE group.

I spent time as an intern at Google (in the Mountain View, CA office in 2011, and the NYC office in 2012), Intel (in the Hudson, MA office in 2014), and Microsoft Research (in the Redmond, WA office in 2015). Before my PhD, I completed my Computer Science B.S. at Indiana University (in my hometown of Bloomington, IN).

Research

I have worked on a variety of computer architecture and systems projects over the course of my PhD, with a general theme of performance and energy efficiency analysis. While interning at Google, I developed performance analysis techniques to improve datacenter scheduling. At Intel, I worked on a project to speed up the process of designing computational GPUs, and at Microsoft Research I helped determine how to efficiently map computer vision applications to emerging memory technologies. At Columbia, I have studied methods for reducing power using operating system and compiler knobs, explored new ways to analyze parallel performance, and developed a language extension for low-energy computing. For more details, see my publications.

Conference Publications

  • M. Kambadur, M. A. Kim, NRG-Loops: Adjusting Power from Within Applications CGO 2016. pdf
  • M. Kambadur, S. Hong, J. Cabral, H. Patil, C.K. Luk, S. Sajid, M. A. Kim, Fast Computational GPU Design with GT-Pin, IISWC 2015. pdf slides
  • M. Kambadur, M. A. Kim, An Experimental Survey of Energy Management Across the Stack, OOPSLA 2014. pdf
  • M. Kambadur, K. Tang, M. A. Kim, ParaShares: Finding the Important Basic Blocks in Multithreaded Programs, Euro-Par 2014. pdf
  • M. Kambadur, T. Moseley, R. Hank, M. A. Kim. Measuring Interference Between Live Datacenter Applications, SC 2012. pdf
  • M. Kambadur, K. Tang, M. A. Kim. Harmony: Collection and Analysis of Parallel Block Vectors, ISCA 2012. pdf

    • Journal Publications

  • M. Kambadur, K. Tang, M. A. Kim. Collection, Analysis, and Uses of Parallel Block Vectors, IEEE Micro Top Picks 2013. pdf

    • Workshops, Technical Reports, Extended Abstracts, and Thesis Work

  • M. Kambadur, Navigating the Hardware/Software Efficiency Gap, Thesis Proposal, November 2014.
  • M. Kambadur, M. A. Kim, Trading Functionality for Power within Applications, APPROX Workshop at PLDI 2014. pdf
  • M. Kambadur, M. A. Kim, Energy Exchanges: Internal Power Oversight for Applications, CUCS Tech Report #009-14. pdf
  • M. Kambadur, K. Tang, J. Lopez, and M. A. Kim. Parallel Scaling Properties from a Basic Block View. SIGMETRICS 2013. pdf poster

    • For fun, here is a wordle containing frequent words from my written projects:

    Wordle

    Personal Interests

    I enjoy traveling, volunteering, tennis, hiking, yoga, and reading all kinds of fiction novels.

    Last edited December 2015