Putting AI to the test
What LLMs really change in reverse-engineering and vulnerability research
WEBINAR | Tuesday, 24th November 2026 | 2 Sessions
Reverse-engineering used to take days or weeks of work and rare expertise. Today, anyone with a monthly AI subscription can attempt it. LLMs are also finding bugs and turning them into exploitable vulnerabilities.
So what can these models really do, and where do they break down?
On Tuesday 24th November, Quarkslab’s researchers share the results of our own benchmarks in two live sessions.
Session 1: Reverse-engineering in the age of AI
11:00 CET · 45 min · Damien Aumaitre, Security Researcher
We benchmarked frontier models on reverse-engineering, then pushed them further: can they understand obfuscation, or even strip it away? Damien shares the results, and what they mean for protecting software, including how we keep our obfuscator QShield a step ahead.
Session 2: A hybrid approach to vulnerability research
14:00 CET · 30 min · Julien Lair, R&D Engineer
Ask an LLM the same question twice and you may get two different answers. In security, that can mean a missed vulnerability. Julien will show how we combine deterministic analysis with model reasoning to find vulnerabilities in source code more reliably and systematically.
Attend one session or both, each followed by a live Q&A. Can’t join live? Register anyway and we’ll send you the recording.
