Presented our paper at RAID 2026 in Lancaster

On 13 October 2026, I presented our paper “Small, Free, and Effective: Orchestrating Open-Weight Small Language Models to Outperform Single LLM for Malware Analysis” at the 29th International Symposium on Research in Attacks, Intrusions and Defenses (RAID 2026) in Lancaster, United Kingdom, as part of the Malware Detection & Analysis session (programme).

The paper asks whether security teams really need large, costly, closed-source LLMs for malware analysis. We show that a carefully orchestrated team of small, open-weight language models, running locally at zero API cost, can match and even outperform a single large LLM, while keeping sensitive samples in-house.

Thank you to the session chair, the RAID organisers, and everyone who came along with great questions and discussion afterwards. This work was done with my supervisors Dr. Budi Arief and Prof. Shujun Li at the University of Kent.

📄 Read the paper: PDF