Matthew Zhang

I'm a fourth-year student currently double majoring in Statistics and Computer Science. I enjoy being able to form hypotheses from messy data and debug complex systems to quantify the root causes of hard problems.

I have both a background in industry—with agent-based validation for in-car voice systems at Lucid Motors and building 0 → 1 full-stack software for research labs that supports real experiments—and an academic research background in networks, specifically network measurement and validation.

I'm also interested in the oracle problem, where validation is only as reliable as the standard used to judge whether a system is correct. I want to build validation methods whose judgments are measurable, auditable, and grounded in real-world outcomes.

If I were to pick one area to work in for the rest of my life, it would be forward-deployed engineering: specifically, being able to understand customers and build a system that achieves their goals within a set of constraints, then validate it statistically to make it not only reliable, but also scalable later. I think it is quite similar to research in the way you turn the unknown into measurable hypotheses that can be iterated on and developed, which I enjoy.

Our DualPI2 for Mahimahi paper is published in ACM SIGCOMM CCR.
Selected research

Work I've authored

All publications →
Second author · ACM SIGCOMM Computer Communication Review

A DualPI2 Module for Mahimahi

An L4S-capable DualPI2 implementation for the Mahimahi emulator, paired with a statistically grounded evaluation of its fidelity to Linux kernel behavior.

Read details →
Experience

Where I've worked

Full experience →
Jun 2026 — Present

Lucid Motors

Software Engineer · In-Vehicle AI Assistant Engineer

Built agent-based validation for Lucid's in-car voice assistant, testing the full path from voice activation through network hops to vehicle signal execution.

Jul 2025 — Jun 2026

Mayer Lab · UC Santa Barbara

Founding Software Engineer

Sole-built Illusion Classroom, a U.S. Navy-funded 0 → 1 full-stack platform that supports daily video-based HCI experiments across UCSB labs.

Writing

From Medium

All writing →