Jayashree Mohan

Department of Computer Science · University of Texas, Austin · jaya@cs.utexas.edu

I am a fifth year PhD candidate in the department of Computer Science at The University of Texas, Austin. I am advised by Prof. Vijay Chidambaram. I also work closely with Amar Phanishayee. My research interests include performance and reliability of file and storage systems in general. More recently, I have been exploring the storage stack and data pipeline in ML training. My thesis focuses on accelerating DNN training from a storage perspective. I am a member of the UT SaSLab and LASR.

Before joining UT, I graduated with a B.Tech in Computer Engineering from National Institute of Technology Karnataka in 2016.

I’m looking for full-time research positions in the industry starting Fall 2021, in India. [CV]

Recent News

  • [Jan 2021] Our work on memory optimization for DNNs is accepted at ICLR'21 (Spotlight).
  • [Dec 2020] Our work on fine-grained DNN checkpointing is accepted at FAST'21.
  • [Dec 2020] Our work on analyzing and mitigating data stalls in DNN training is accepted at VLDB'21.
  • [July 2020] Honored to be awarded the UT Dean's Prestigious Fellowship Supplement.
  • [July 2020] Passed my PhD thesis proposal. Officially a Doctoral Candidate.
  • [Dec 2019] Video of my talk on CrashMonkey @Heisenbug19 is now online at video.
  • [Oct 2019] Invited to present CrashMonkey at Heisenbug 2019 Moscow.
  • [Aug 2019] Presented our work on GDPR Dark Patterns at Poly'19.
  • [July 2019] Our RECIPE to convert DRAM indexes to Persistent Memory indexes is accepted at SOSP'19!
  • [Jan 2019] Honored to be selected as a Microsoft PhD Fellow!

Education

University of Texas at Austin

Ph.D.
Computer Science

GPA: 3.91

2016 - 2021(expected)

National Institute of Technology Karnataka

B.Tech.
Computer Engineering

GPA: 9.89/10

2012-2016

Publications

Conference Publications

  1. CheckFreq: Frequent, Fine-Grained DNN Checkpointing FAST 2021
    Jayashree Mohan, Amar Phanishayee, Vijay Chidambaram
    [pdf] [slides] [bibtex]

  2. Analyzing and Mitigating Data Stalls in DNN Training VLDB 2021
    Jayashree Mohan, Amar Phanishayee, Ashish Raniwala, Vijay Chidambaram
    [pdf] [bibtex] [extended version]

  3. Memory Optimization for Deep Networks ICLR 2021
    Aashaka Shah, Chao-Yuan Wu, Jayashree Mohan, Vijay Chidambaram, Philipp Krahenbuhl
    [pdf]

  4. RECIPE : Converting Concurrent DRAM Indexes to Persistent-Memory Indexes SOSP 2019
    Se Kwon Lee, Jayashree Mohan, Sanidhya Kashyap, Taesoo Kim, Vijay Chidambaram
    [pdf] [slides] [bibtex] [extended version]

  5. Instalytics: Cluster Filesystem Co-design for Big-data Analytics FAST 2019
    Muthian Sivathanu, Midhul Vuppalapati, Bhargav Gulavani, Kaushik Rajan, Jyoti Leeka, Jayashree Mohan, Piyus Kedia
    [pdf] [slides] [bibtex]

  6. Finding Crash-Consistency Bugs with Bounded Black-Box Crash Testing OSDI 2018
    Jayashree Mohan, Ashlie Martinez, Soujanya Ponnapalli, Pandian Raju, Vijay Chidambaram
    [pdf] [slides] [bibtex] [full video] [extended version] [code] [blog post] [demo video]

  7. A Reinforcement Learning Approach to Optimize Downloads Over Mobile Networks COMSNETS 2017
    Jayashree Mohan, Angad Vittal, K Chandrasekaran, Bhaskar Krishnamachari
    [pdf] [bibtex]

  8. Optimizing Downloads over Random Duration Links in Mobile Networks ICCCN 2016
    Amber Bhargava, Spencer Congero, Timothy Ferrell, Alex Jones, Leo Linsky, Jayashree Mohan, Bhaskar Krishnamachari
    [pdf] [bibtex]

  9. Reducing DNS Cache Poisoning Attacks ICACCS 2015
    Jayashree Mohan, Shruthi Puranik, K Chandrasekaran
    [pdf] [bibtex]
    Best Paper Award!

Workshop Publications

  1. Analyzing GDPR Compliance Through the Lens of Privacy Policy Poly 2019
    Jayashree Mohan, Melissa Wasserman, Vijay Chidambaram
    [pdf] [slides] [bibtex]

  2. Storage on Your Smartphone Uses More Energy Than You Think HotStorage 2017
    Jayashree Mohan, Dhathri Purohith, Matt Halpern, Vijay Chidambaram, Vijay Janapa Reddi
    [pdf] [slides] [bibtex]

  3. The Dangers and Complexities of SQLite Benchmarking ApSys 2017
    Dhathri Purohith, Jayashree Mohan, Vijay Chidambaram
    [pdf] [slides] [bibtex]

Journal Publications

  1. Instalytics: Cluster Filesystem Co-design for Big-data Analytics TOS 2020
    Muthian Sivathanu, Midhul Vuppalapati, Bhargav Gulavani, Kaushik Rajan, Jyoti Leeka, Jayashree Mohan, Piyus Kedia
    [pdf] [bibtex]

  1. CrashMonkey and ACE: Systematically Testing File-System Crash Consistency TOS 2019
    Jayashree Mohan, Ashlie Martinez, Soujanya Ponnapalli, Pandian Raju, Vijay Chidambaram
    [pdf] [bibtex]

Work in Progress

  1. Analyzing IO Amplification in Linux File Systems ApSys 2017
    Jayashree Mohan, Rohan Kadekodi, Vijay Chidambaram
    [poster abstract] [arxiv preprint]
    Best Poster Award!

Awards

  • UT Austin Dean's Prestigious Fellowship Supplement 2020
  • Microsoft Research PhD Fellowship 2019-2021
  • UT Austin Dean's Prestigious Fellowship Supplement 2019
  • Best Poster Award at ApSys 2017
  • Department Gold Medalist in B.Tech 2016
  • Institution of Engineers (IE) Certificate for highest GPA in CS, for all 4 years of undergraduate study. 2016
  • Viterbi-India Research Fellowship 2015
  • IIT Delhi Summer Research Fellowship 2014
  • Best Paper Award at ICACCS 2015
  • Ministry of Human Resource Development (MHRD India) Scholarship 2012-2015

Experience

Microsoft Research Redmond
Position : Research Intern
Mentor : Amar Phanishayee
Project : As a part of Project Fiddle, I investigated the data pipeline of Deep Neural Network training. Our work focused on identifying and mitigating disk and CPU stalls.
May 2020 - Aug 2020
May 2019 - Aug 2019
Microsoft Research Cambridge, UK
Position : Research Intern
Mentor : Andromachi Chatzieleftheriou, Richard Black, Ant Rowstron
Project : Worked on analyzing the design space and modelling the write head in Project Silica. This project explores the use of quartz glass as the mass storage media for cloud.
June 2018 - Aug 2018
Microsoft Research Bangalore
Position : Research Intern
Mentor : Muthian Sivathanu
Project : At MSR India, I contributed to Project Instalytics, that co-designs the compute and storage layers for efficient big data analytics. As a part of this project, I worked on imparting semantic knowledge to a distributed storage system.
May 2017 - Aug 2017

Fun Stuff

Apart from being a Systems Researcher, I enjoy painting, arts, and craft. My favourite medium is acrylic on canvas. I also make stuffed toys. Here's a preview :)


Adapted from template