Hello, I'm Arash Sarabi

Ph.D. Candidate in Computer Science

Computer Systems Design & Implementation | Distributed Systems | Systems for Machine Learning | Hardware Acceleration

View My Work Contact Me

About Me

Arash Sarabi

Ph.D. candidate in Computer Science at Arizona State University, ranked 1st in my cohort with a perfect 4.0 GPA. Recipient of the Outstanding Thesis Award and IEEE-HKN Honor Society member. Expected graduation: May 2026.

My research focuses on Computer Systems Design & Implementation, spanning distributed systems, network systems, systems for machine learning, and hardware acceleration. I designed, implemented, and deployed the MLED distributed system framework on the FABRIC national testbed across 30+ geographically distributed sites, achieving a 14,000× reduction in undetected errors for petabyte-scale data transfers.

My work includes cloud-scale systems performance optimization, FPGA-based hardware acceleration (50× speedup in checksum computation), and building ML infrastructure for TB-scale datasets. I have published in POMACS (ACM SIGMETRICS) and IEEE Communications Letters, with presentations at ACM SIGMETRICS 2025 and IEEE FCCM 2025. I also co-founded Expeditise LLC, an AI startup building intelligent tools for everyday workflows.

Ph.D. in Computer Science, Arizona State University
Research Assistant & Co-Founder at Expeditise
Outstanding Thesis Award | IEEE-HKN Honor Society
Tempe, AZ

Education

Ph.D. in Computer Science

Arizona State University (ASU), Tempe, AZ

Aug 2021 – May 2026 (Expected)

Dissertation: "Multi-Level Error Detection (MLED) for Petabyte-Scale Reliable Data Transfers"

Advisor: Dr. Violet R. Syrotiuk

GPA: 4.0/4.0 | Outstanding Thesis Award | Ranked 1st in Ph.D. Cohort | IEEE-HKN Honor Society

M.Sc. in Computer & Information Technology Engineering

Sharif University of Technology (SUT)

GPA: 17.75/20 (3.90/4.0 WES)

B.Sc. in Computer & Information Technology Engineering

Amirkabir University of Technology (AUT)

GPA: 17.27/20 (3.72/4.0 WES)

Ranked 2nd among 50 students

Experience

Research Assistant

Aug 2021 – Present

Arizona State University | Tempe, Arizona

  • Experimental Distributed Systems Research: Designed, implemented, and deployed the MLED distributed system framework on the FABRIC national testbed spanning 30+ geographically distributed sites. Built end-to-end prototype demonstrating 14,000× reduction in undetected errors for petabyte-scale data transfers across wide-area networks.
  • Cloud-Scale Systems Performance: Evaluated system performance across distributed cloud infrastructure using the FABRIC testbed; optimized network protocols for multi-datacenter data transfers; conducted micro-architectural bottleneck analysis of packet processing pipelines in cloud environments.
  • Hardware-Software Co-Design and Acceleration: Integrated FPGA-based acceleration for network protocol processing; designed algorithms leveraging hardware accelerator capabilities; achieved 50× speedup in checksum computation. Presented research on "Using FPGAs to Accelerate Error Detection in MLED" at IEEE FCCM 2025.
  • Systems for Machine Learning: Built infrastructure for deploying containerized ML workloads on Kubernetes clusters; optimized distributed ML training pipelines for TB-scale datasets; automated Google Cloud Platform infrastructure with Terraform. Developed Graph Neural Network architecture for spatiotemporal prediction.
  • Network Systems Design: Designed optimal configuration algorithms for distributed data transfer systems; developed network partitioning strategies using mathematical optimization. Created formal models characterizing undetected error probability as a function of file size, network conditions, and system policy configurations.

Co-Founder

Jan 2024 – Present · 2 yrs 1 mo

Expeditise LLC · Full-time | Tempe, Arizona, United States · On-site

  • Architected scalable distributed backend infrastructure with microservices architecture, load balancing, and fault tolerance mechanisms. Built containerized deployments with automated CI/CD pipelines and infrastructure-as-code solutions using Terraform and Ansible.
  • Leading cross-functional team of 18 developers, establishing development workflows and coordinating system architecture decisions balancing performance, scalability, and maintainability.

My Skills

Programming Languages

Python C/C++ Java Rust JavaScript/TypeScript Golang

Systems & Infrastructure

Linux/Unix Docker & Kubernetes Terraform & Ansible Google Cloud Platform Distributed Systems Design

Machine Learning & Performance

PyTorch & TensorFlow Graph Neural Networks Linux perf & Benchmarking FPGA Acceleration Hardware-Software Co-Design

Publications & Awards

Refereed Journal Papers

POMACS 2025

Design and Modeling of a New File Transfer Architecture to Reduce Undetected Errors Evaluated in the FABRIC Testbed

P. Jain, A. Sarabi, A. Matta, V.R. Syrotiuk

Proceedings of the ACM on Measurement and Analysis of Computing Systems (POMACS - Journal of ACM SIGMETRICS), 9(2):1-42, May 2025

IEEE 2021

MLED: A Layered Architecture for Reducing Undetected Error Probability in File Transfer

A. Sarabi, I. Matta, V.R. Syrotiuk

IEEE Communications Letters, 25(11):3542-3545, Aug 2021

arXiv 2025

Forecasting Coccidioidomycosis (Valley Fever) in Arizona: A Graph Neural Network Approach

A. Sarabi, A. Sarabi, H. Yan, B. Sterner, P. Jevtić

arXiv:2507.10014, July 2025

Conference Presentations & Extended Abstracts

SIGMETRICS 2025

Design and Modeling of a New File Transfer Architecture to Reduce Undetected Errors

ACM SIGMETRICS 2025 (Extended Abstract), Stony Brook, NY, June 2025

FCCM 2025

Using FPGAs to Accelerate Error Detection in MLED

IEEE FCCM 2025 (Invited Talk - OCT Workshop), Fayetteville, AR, May 2025

KNIT 2023

Large-Scale File Transfer Optimization Strategies

KNIT 7: FABRIC Community Workshop, 2023. Best Demo & Poster Award

Patent

Patent 2025

Recursive File Transfer Architecture to Reduce Undetected Errors

U.S. Provisional Patent Application No. 63/864,142, Aug 2025

Awards & Honors

Outstanding Thesis Award

ASU Graduate College Graduate Research Support Program (GRSP), Sep 2024

Research Award

Graduate Student Government (GSG), Nov 2023

Best Demo & Poster Award

KNIT 7: A FABRIC Community Workshop, Sep 2023

IEEE-HKN Honor Society

IEEE-Eta Kappa Nu, 2022-2025

Ranked 1st in Ph.D. Cohort

Arizona State University, 2021-Present

Service & Leadership

Peer Reviewer (2019 – Present): 19 reviews of 18 manuscripts for IEEE INFOCOM, Computer Networks (Elsevier), IEEE WCNC, Joint European Conference on Networks and Communications, 6G Summit

Web of Science Verified Reviewer: ResearcherID: ABG-2912-2021, ORCiD: 0000-0003-3611-4459

Session Host: CNERT 2024, 2023, 2021 (in conjunction with IEEE INFOCOM)

Teaching Assistant, Arizona State University (2021-2025): Computer Networks, Data Structures & Algorithms, Foundations of Algorithms, Data Visualization, Software Project Management, Principles of Programming Languages

My Projects

MLED Architecture Diagram

MLED: Multi-Level Error Detection Architecture

Designed and deployed distributed system on FABRIC testbed spanning 30+ sites. Achieved 14,000× reduction in undetected errors for petabyte-scale data transfers. Published in POMACS (ACM SIGMETRICS) and IEEE Communications Letters. U.S. Patent Pending.

Distributed Systems FABRIC Testbed Network Protocols POMACS
Learn More

FPGA-Accelerated Error Detection

Integrated FPGA-based acceleration for network protocol processing, achieving 50× speedup in checksum computation. Presented at IEEE FCCM 2025 (premier hardware/FPGA venue).

FPGA Hardware Acceleration IEEE FCCM

Forecasting Valley Fever in Arizona: A Graph Neural Network Approach

Developed the first GNN model for forecasting Coccidioidomycosis incidence in Arizona. Integrates surveillance data with environmental predictors including soil conditions, atmospheric variables, and air quality metrics. Published on arXiv (July 2025).

Graph Neural Networks Machine Learning arXiv 2025
View on arXiv

Kubernetes-based ML Infrastructure

Built infrastructure for deploying containerized ML workloads on Kubernetes clusters. Optimized distributed ML training pipelines for TB-scale datasets with Google Cloud Platform automation using Terraform.

Kubernetes ML Infrastructure GCP Terraform

Expeditise: AI Solutions for Everyday Problems

Co-founded AI startup building intelligent tools that seamlessly integrate into daily workflows. Developing ExpressReply, an AI-powered smart keyboard with context-aware suggestions, OCR reading, and privacy-first design. Processing 100K+ daily requests with <100ms latency.

AI/ML Startup GPT API React Native Cloud Services
Visit Website

Get In Touch

Phone

+1-480-593-2454

Location

Tempe, AZ

CV Information

Hi! I'm Arash's CV assistant. Ask me about his impressive skills, education, or projects. For detailed information, you can contact him directly!