Tarek Elsayed

Tarek Elsayed

AI Safety Researcher · Systems Engineer

Vancouver, Canada

I’m a software engineer at Amazon working on internal AI systems and workflows. Previously, I was a research intern at Microsoft Research. I hold an MSc in Computing Science from Simon Fraser University.

I’m interested in AI safety, particularly evaluation and security, and in systems engineering with Rust. My work examines benchmark integrity and contamination, and how to measure model capabilities on software and security tasks.

I also participated in BlueDot Impact’s Technical AI Safety cohort.

Research interests: AI evaluation · AI security · agent safety · benchmark integrity · systems & Rust

Publications

RustMizan: A Compilable, Contamination-Aware Benchmarking Framework for Rust Vulnerabilities

Tarek Elsayed, Shiping Yang, Eunsong Koh, Sanika Goyal, Vincent Huang, Paul Ngo, Nathan Young, Mohammad Omidvar Tehrani, Alvyn Kang, Arnell Kang, Zeyu Chen, Angélica Moreira, Xuan Feng, Angel X. Chang, Nick Sumner, Steven Y. Ko

NeurIPS 2026 Evaluations and Datasets Track

A framework for evaluating LLM agents on real-world Rust vulnerabilities using compilable programs and paired code mutations to probe contamination and robustness.

Selected experiences

Amazon 2026–present
Software Development Engineer

Internal agentic AI systems

Microsoft Research 2026
Research Intern

Led SysEduBench. I evaluated LLMs on systems knowledge and implementation tasks with Inspect AI.

MLH with Meta 2024
Software Engineering Fellow

Worked on rust-analyzer’s code assists and editor configuration.

Google Summer of Code 2022–24
Contributor & Mentor · mlpack

Implemented DDPG and TD3 agents and ML examples in mlpack (2022–23). Mentored contributors in 2024.

Linux Foundation 2023–24
Open Source · LFX Mentorship

Worked on Solang, a compiler written in Rust for multiple targets.