Siqi Li

Hi! I’m Siqi 👋

I work on AI robustness and alignment, with a particular interest in what happens when intelligent systems leave clean benchmarks and enter the real world.

My research asks a simple but uncomfortable question:
how do we know an AI system is doing the right thing when it fails quietly, faces distribution shift, or is intentionally attacked?

While much of robustness research is mathematical, I approach the problem both theoretically and engineering-first. I build systems, stress them, break them, and then ask what guarantees actually survive deployment.

I’m a PhD student in EECS at UC Irvine, advised by
Prof. Yasser Shoukry in the
Resilient Cyber-Physical Systems Lab.
Previously, I was a visiting researcher at Caltech, working with
Prof. John Doyle on language-to-robot control.

How I Think About Robustness

I’m especially interested in robustness for vision-language-action systems, where failures are subtle, delayed, and hard to detect.

My work spans:

• Formal and provable guarantees (e.g., adversarial detection with correctness proofs)
• System-level verification for VLA-based robot policies
• Engineering-heavy evaluation pipelines that expose real failure modes

I care less about making models look good on paper — and more about making failure visible, interpretable, and actionable.

Building Real Systems (and Breaking Them)

Beyond research prototypes, I enjoy building interactive systems that force models to operate under realistic constraints.

I’ve worked on:

• Language-to-robot control systems that integrate learning, planning, and feedback
• VR and game-like environments (including Meta Quest–style setups) to study perception, interaction, and failure in embodied agents
• Simulation-driven stress tests that reveal where “robust” models actually collapse

Game engines and VR are especially useful here — they let us design controlled worlds where failures are unavoidable, observable, and repeatable.

A Bit More About Me 🌍

I work fluently in three languages (Mandarin, English, and German), both conversationally and academically, and I’ve lived and studied across Asia, Europe, and the US.

I enjoy travel, food, and building things that work — which is probably why I’m drawn to problems where theory meets reality.

I believe the most interesting AI research is:

• rigorous but not fragile
• principled but hands-on
• serious, yet a little fun

If that resonates, feel free to reach out — I’m always happy to chat.