Ayush Goel

Ayush Goel

Systems Research Scientist

Meta

Biography

Ayush Goel is a Research Scientist at Meta in the Network.AI group. He previously worked at industry research labs like Hewlett Packard Labs, IBM Research. He got his PhD in Computer Science from the University of Michigan in Oct 2023. He was affiliated with the Systems Lab where he was advised by Harsha V. Madhyastha and Ravi Netravali. His research interests lie at the intersection of distributed systems and programming languages. During his PhD, he investigated the application of PL techniques on the modern web, and built systems that leverage fine-grained data-flow, control-flow analysis of web computations to improve information exchange on the web. He has also worked on geo-distributed storage systems and software-defined WANs.

Interests
  • High performance computing
  • Networked systems for AI
  • Program analysis
Education
  • PhD in Computer Science, 2023

    University of Michigan

  • Msc in Computer Science, 2019

    University of Michigan

  • BTech in Computer Engineering, 2016

    IIIT Delhi

Experience

Research Scientist Meta SF Bay Area, California
April 2025 – Present
  • Research Scientist in the Network.AI group in the NetInfra org
Research Scientist Hewlett Packard Labs SF Bay Area, California
November 2023 – April 2025
  • Research Scientist in the Networking and Distributed Systems Lab
  • Leading efforts on distributed systems for LLM serving
  • Other areas: Network for AI, CXL-based disaggregated memory solutions
Research Engineer IBM Research New Delhi, India
July 2016 – August 2017
  • High performance computing
  • Built orchestration tools for Machine-learning-as-a-Service for IBM Watson

Recent News
(All News)

November 2025 Two patents, based on my work at HPE labs, on optimized LLM inference serving got filed.
October 2025 Our work on Workload-Adaptive Cluster Scheduler for Latency-Optimal LLM Inference Serving got accepted at EuroSys'26.
April 2025 I am thrilled to join Meta’s Network Infra for AI group (Network.AI) as a Research Scientist.
February 2025 Excited to serve on the NSDI'26 technical program committee.
February 2024 Excited to serve on the NSDI'25 program committee.

Featured Publications

Quickly discover relevant content by filtering publications.
(2024). Sprinter: Speeding up High-Fidelity Crawling of the Modern Web. In NSDI'24.

PDF Code Dataset

(2022). Making Links on Your Web Pages Last Longer Than You. In HotNets'22.

PDF

(2022). Jawa: Web Archival in the Era of JavaScript. In OSDI'22.

PDF Code Dataset Poster Slides Video Source Document

(2021). Horcrux: Automatic JavaScript Parallelism for Resource-Efficient Web Computation. In OSDI'21.

PDF Code Slides

(2021). Rethinking Client-Side Caching for the Mobile Web. In HotMobile'21.

PDF Code Video