Sohaib Ahmad

I am a Research Scientist at Meta working in the Facebook Stories Infrastructure team to improve the efficiency of ML model serving.

I recently completed my Ph.D. at the University of Massachusetts, Amherst working under the supervision of Prof. Ramesh Sitaraman and Prof. Hui Guan. My research interests lie at the intersection of systems and machine learning, particularly improving the resource efficiency of ML inference and model serving. As the demand for ML grows at a much faster rate than the underlying infrastructure to host and serve it, my research aims to improve the efficiency of existing infrastructure to serve this growing demand, through intelligent resource allocation and workload scheduling.

During my Ph.D., I interned at Meta with their Serverless Computing team. I also interned with Bell Labs research twice to work on scheduling and resource allocation for ML model serving and training on the cloud and edge.

Before starting my Ph.D., I completed my undergrad from LUMS in 2017 under the supervision of Prof. Ihsan Ayyub Qazi where my research focused on computer networks and measurement.

News

  • [Feb 2025] Our paper, “DiffServe: Efficiently Serving Text-to-Image Diffusion Models with Query-Aware Model Scaling” has been accepted to MLSys 2025!
  • [Mar 2024] Our paper, “Loki: A System for Serving ML Inference Pipelines with Hardware and Accuracy Scaling” has been accepted to HPDC 2024!
  • [Aug 2023] Our paper, “Proteus: A High-Throughput Inference-Serving System with Accuracy Scaling” has been accepted to ASPLOS 2024!
  • [May 2023] 🏆 Our paper, AggFirstJoin, won the Best Paper Award at ACM/IEEE CCGrid 2023!

Awards & Service

  • [2023] 🏆 Best paper award at ACM/IEEE CCGrid
  • [2023] Received the Manning College Dissertation Writing Fellowship
  • [2022] Shadow Program Committee Member for EuroSys
  • [2022] Panelist at CS Research Night at UMass
  • [2018] 🏆 Best paper of the year award at ACM SIGCOMM CCR
  • [2017] Awarded the Krithi Ramamritham scholarship at UMass Amherst for outstanding incoming student in sytems research