Rajmohan Rajaraman
Northeastern University, MA·Computer Science
Large computations, such as training and running AI models, are increasingly split across many machines in cloud systems and datacenters. Deciding which machine runs which task, and when, is hard.
- Fields
Computer Science
Mathematics & Statistics
Engineering
- Topics
scheduling algorithms
online algorithms
approximation algorithms
- Grants
- 1 known grant, through Jan 2027
What they work on
- Design communication-aware algorithms for scheduling precedence-constrained jobs, which are tasks that depend on earlier tasks. The jobs run on networks of machines with differing speeds, capabilities and delays, and the algorithms use new linear-programming techniques.
- Develop scheduling methods for splittable jobs on reconfigurable machines, and minimum-stretch embeddings that map computation graphs onto hardware networks while keeping communicating tasks close.
- Create online algorithms for migrating processes and servers as requests arrive unpredictably. These include heterogeneous variants of the k-server problem and online balanced graph partitioning.
- Study learning-augmented algorithms that use predictions of processing times and communication needs, and extend these frameworks to reconfigurable architectures and mobile ad hoc networks.
More about the research
Machines differ in speed and capability. Tasks depend on each other's results and must exchange data over the network. New requests also keep arriving without warning. This project designs algorithms that schedule and move computations efficiently while explicitly accounting for communication costs, the goal being faster processing of large AI workloads and better resource allocation in cloud systems.
Why it matters
Scheduling and placement decisions determine how fast large AI models and cloud workloads run and how much hardware they tie up. Algorithms that account for communication between distributed computations could help cloud providers and datacenter operators run large AI tasks more efficiently. They could also improve performance for business and mission-critical systems.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Bijaya Adhikari
University of Iowa, IA·Computer Science·Early-career
Formulate a new class of online optimization problems for healthcare-associated infection (HAI) interventions and design online algorithms, which decide as data arrive, with worst-case approximation and hardness guarantees.
Arvind Krishnamurthy
University of Washington, WA·Computer Science
Design NetFusion, a rack-level pool of programmable network interface cards (SmartNICs) that perform packet and network-function processing for many servers, with safe and fair sharing among applications.
Rajan Udwani
University of California-Berkeley, CA·Engineering·Early-career
Formulate new models of online resource allocation that capture the complex features, objectives and constraints of applications such as shared mobility, electric-vehicle battery swapping and online advertising.
The school
Northeastern University: 27 labs on PI Pages, $347M a year on research, 251 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.