profile_picture
Nitika Saran
Ph.D. Candidate,  Cornell University
nsaran [at] cs [dot] cornell [dot] edu

I am a Ph.D. candidate in computer science at Cornell University, where I am advised by Hakim Weatherspoon. My research interests are broadly in systems and networking, with a focus on networking infrastructure for AI and cloud workloads. My recent work proposes new paradigms for datacenter networking that use optical circuit switches to reduce cost and power consumption compared to existing packet-switched designs.

During my Ph.D., I’ve been fortunate to intern at several wonderful research labs. In summer 2025, I worked with Paolo Costa and others at Microsoft Research Cambridge, towards efficient network design for LLM training and inference. From 2022-23, I worked with Sylvia Ratnasamy and the Google Systems Research Group on traffic engineering for software-defined wide area networks.

Before Cornell, I was a research fellow at Microsoft Research India, working with Muthian Sivathanu and Ram Ramjee. Here, I worked on Varuna, a large-scale distributed training framework for LLMs like GPT and BERT. Earlier, I spent a year in Microsoft’s engineering division earned my undergraduate degree from IIIT Delhi.

Research Interests

  • Datacenter Networking
  • Systems for ML
  • Distributed Systems

Education

Cornell University
2021 - 2026 (expected)
Ph.D. in Computer Science
IIIT Delhi
2014 - 2018
B.Tech in Computer Science

Publications

  • Semi-Oblivious Reconfigurable Datacenter Networks
    Nitika Saran, Daniel Amir, Tegan Wilson, Robert Kleinberg, Vishal Shrivastav, Hakim Weatherspoon
    HotNets 2024
  • dSDN: A Decentralized SDN Architecture for the WAN
    Alex Krentsel, Nitika Saran, Bikash Koley, Subhasree Mandal, Ashok Narayanan, Sylvia Ratnasamy, Ali Al-Shabibi, Anees Shaikh, Rob Shakir, Ankit Singla, Hakim Weatherspoon
    SIGCOMM 2024
  • Shale: A Practical, Scalable Oblivious Reconfigurable Network
    Daniel Amir, Nitika Saran, Tegan Wilson, Robert Kleinberg, Vishal Shrivastav, Hakim Weatherspoon
    SIGCOMM 2024
  • Breaking the VLB Barrier: Improving Oblivious Reconfigurable Networks with High Probability
    Tegan Wilson, Daniel Amir, Nitika Saran, Robert Kleinberg, Vishal Shrivastav, Hakim Weatherspoon
    STOC 2024
  • Varuna: Scalable, Low-cost Training of Massive Deep Learning Models
    Nitika Saran*, Sanjith Athlur*, Muthian Sivathanu, Ram Ramjee, Nipun Kwatra
    EuroSys 2022
    won Best Paper Award!

Teaching Experience