SYNOPSIS

mrkatebzadeh [--experience] [--projects] [--publications]

DESCRIPTION

Optimizing large language model (LLM) infrastructure while exploring distributed systems, replication protocols, and performance predictability in datacenter networks.

Past contributions include RISC-V on FPGA and systems projects within high-performance computing; hobbies span photography, chess, manga, and anime.

EXPERIENCE

Research Scientist, Huawei Technologies Research & Development (2026—)

Postdoctoral Research Associate, The University of Edinburgh (2024–2025)

CPU Architecture Intern, Huawei Technologies Research & Development (2023)

Postgraduate Researcher (PhD), The University of Edinburgh (2017–2022)

PROJECTS

Chaapar — cache-coherence inspired replication logic on Arm SmartNICs.

Saba — application-aware bandwidth allocation for datacenter workloads.

RPerf — precise RDMA switch measurement toolkit.

Hermes — fast, fault-tolerant and linearizable RDMA protocol.

Deterministic QEMU — tooling for eliminating non-determinism in QEMU.

BrainF Compiler — optimized BrainF compiler written in Rust.

Bigkernel — CPU-GPU pipelining for big data using LLVM.

Darya — terminal disk usage navigator for headless systems.

Animestan — TUI for searching and playing anime content.

Cheshmak — context-aware prompt hook for git project summaries.

FTwitter — TUI API for Twitter data manipulation.

PUBLICATIONS

Reliable Replication Protocols on SmartNICs — PaPoC / 2025

Saba: Rethinking Datacenter Network Allocation — EuroSys / 2023

Hermes: A Fast, Fault-Tolerant and Linearizable Replication Protocol — ASPLOS / 2020

Evaluation of an InfiniBand Switch — ISPASS / 2020

Bankrupt covert channel — WOOT / 2020

SUPPORTED EXTENSIONS

*.rs *.c *.cpp *.py *.java *.js *.sh

ENVIRONMENT VARIABLES

LANG en_US.utf-8 · fa_IR.utf-8

CONFIG

dwm, wezterm, emacs, nushell.

SEE ALSO

email(4), scholar(4), github(4), linkedin(4), x(4).

KNOWN BUGS

Conserving energy.