Diptyaroop Maji

Diptyaroop Maji

AI systems · Distributed systems · Applied ML · Sustainable computing

I am a PhD candidate in the College of Information and Computer Sciences at UMass Amherst, advised by Prof. Ramesh K. Sitaraman. I also work closely with Prof. Prashant Shenoy.

My research spans AI systems and distributed computing, with a focus on sustainability. I develop machine learning methods to forecast the carbon footprint of electricity grids and real-world systems, and design systems that use these signals to guide operational decisions.

I am also currently a Student Researcher at Google's Systems Research Group , where I am working on scheduling ML workloads across geo-distributed clusters. Previously, I worked at VMware Research (internship), IIT Bombay, and Samsung Research Institute Bangalore.

Expected PhD graduation: May 2027.

Research

I work on both systems for machine learning and machine learning for systems. My current research directions include:

  • Systems for AI. Scheduling ML workloads across geographically distributed clusters to reduce emissions while maintaining resource occupancy and meeting user constraints.
  • Machine learning for systems. Using time series foundation models and forecasting methods to support infrastructure decisions, including CarbonX and CarbonCast.
  • Sustainable distributed infrastructure. Designing carbon-aware application delivery and studying how carbon accounting and renewable energy availability affect system optimization.

My broader systems background includes high-performance 5G dataplanes and ongoing work on neural loss recovery for volumetric videoconferencing.

Publications & Manuscripts

Google Scholar
In review
Under review

ReVo: A Cross-Layer Neural Loss Recovery for Reliable Volumetric Videoconferencing

Ankur Aditya*, Diptyaroop Maji* et al.

* Equal student contribution. Manuscripts under review are labeled above.

Research & Work Experience

May 2026 — Present

Google

Student Researcher · Systems Research Group

Developing a scheduler for carbon-aware placement of ML workloads across geo-distributed Google clusters, minimizing system emissions while maintaining resource occupancy and user-specified constraints.

Mentor: Yawen Wang

Aug 2021 — Present

UMass Amherst

Graduate Research Assistant

Building open-source carbon intensity forecasting tools and systems that use these forecasts to reduce emissions at scale, as part of NSF CoDec: Computational Decarbonization of Societal Infrastructures at Mesoscales.

May — Aug 2022

VMware Research

Research Intern

Developed a carbon-aware prototype for the NSX Avi Global Load Balancer. Routing requests using carbon intensity and latency lowered average operational emissions by 21% relative to the native load balancer.

Mentor: Ben Pfaff

Jul 2020 — Jun 2021

IIT Bombay

Project Engineer

Developed a high-performance DPDK-based 5G User Plane Function dataplane, achieving 40 Gbps line rate while enforcing QoS rules for 65K sessions.

PI: Mythili Vutukuru

Jun 2016 — Jun 2018

Samsung Research Institute Bangalore

Software Engineer → Senior Software Engineer

Built and maintained Android Mobile Hotspot features for flagship devices, with production code in the Android framework and UX layers.

Software Engineer: Jun 2016–Feb 2018 · Senior Software Engineer: Mar–Jun 2018

May — Jul 2015

Samsung Research Institute Bangalore

Student Trainee

Developed a file-sharing application using simultaneous Hotspot and Wi-Fi connections across 2.4 and 5 GHz bands for faster transfers between devices.

Education

2021 — Present

UMass Amherst

PhD Candidate · Computer Science

College of Information and Computer Sciences
Advisor: Ramesh K. Sitaraman

GPA 3.90/4 · Expected May 2027

2018 — 2020

IIT Bombay

M.Tech · Computer Science & Engineering

Thesis: Design, Development and Optimization of the User Plane Function (UPF) Dataplane in 5G
Advisor: Mythili Vutukuru

GPA 9.70/10

Read thesis

2012 — 2016

Jadavpur University

B.E. · Computer Science & Engineering

First-Class with Honors
Kolkata, India

GPA 8.47/10

Awards & Honors

  • 2025Best Paper Finalist, ACM e-Energy
  • 2024Best Paper Runner-Up, ACM e-Energy
  • 2024ACM SIGEnergy Student Travel Grant
  • 2019Teaching Assistant of the Month, IIT Bombay
  • 2018All India Rank 42 in GATE, among 107,893 candidates
  • 2016First-Class with Honors, Jadavpur University

Academic Service

  • Chair & organizing committeeCarbonMetrics ’26

    Workshop on Measurements, Modeling, and Metrics for Carbon-Aware Computing · SIGMETRICS

  • Program committeeFMSust ’26

    Workshop on Foundation Models for Energy-Efficient Buildings, Cities, Transportation, and Sustainability · BuildSys

  • Shadow program committeeEuroSys ’26

    21st European Conference on Computer Systems

Teaching

UMass Amherst · Spring 2026

CS 590AE: Wireless and Mobile Networks

Teaching Assistant · Jim Kurose

IIT Bombay · 2018–2020

Teaching Assistant

  • CS 695: Topics in Virtualization & Cloud Computing Spring 2020 · Mythili Vutukuru
  • CS 744: Design and Engineering of Computing Systems Fall 2019 · Umesh Bellur
  • CS 224: Computer Networks Spring 2019 · Kameswari Chebrolu
  • CS 305: Computer Architecture Fall 2018 · Bernard Menezes

Mentorship

2024–2025 · UMass

Tanush Savadi (UG, UMass Amherst)

Undergraduate independent study

Summer 2024 · NSF REU

Jason Chen (UG, Boston University)

Research Experience for Undergraduates

Spring 2023 · UMass

Rohan Lekhwani (MS, UMass Amherst)

Master’s independent study

Media Coverage

Technical background

Programming
C/C++, Java, Python
Systems & networking
DPDK, Docker, AWS, Kubernetes
Machine learning
PyTorch, TensorFlow, Keras

Outside research

I enjoy football, reading, travel, and chess. I follow the Premier League and am always happy to talk football.

Me trying to juggle at home.

Contact

dmaji@cs.umass.edu · dmaji@umass.edu