L2M – Gen AI powered cyber threat assessment platform

We are developing a next-generation cybersecurity tool that uses large language models (LLMs), a form of generative AI, to help small and medium-sized businesses (SMBs) detect hidden weaknesses in their systems before they can be exploited. Unlike traditional tools that rely on fixed rules and often flood users with too many alerts, our platform understands patterns in system logs the way a security expert would. It connects directly to existing log data sources like Zeek, Suricata, Windows Events, or AWS CloudTrail, and uses an AI model combined with the latest threat databases (like CVEs and CISA KEV) to spot outdated technologies or risky configurations—such as old encryption or legacy file-sharing systems. When it finds a risk, the system explains the issue clearly and provides a ready-to-use report, making it easy for non-experts to take action. The platform runs entirely within the organization’s own infrastructure, keeping all sensitive data private, and only shares anonymized updates, making it especially suitable for privacy-sensitive industries like healthcare or utilities. The partner organization will benefit from a cutting-edge solution that enhances cybersecurity without the need for extra staffing or major infrastructure changes. It reduces the time and effort needed for vulnerability assessments, helps prioritize what to fix first, and keeps businesses protected from modern AI-driven threats—while being cost-effective and easy to integrate.

Faculty Supervisor:

Nashid Shahriar

Student:

Partner:

North Forge

Discipline:

Computer science

Sector:

Education; Management of companies and enterprises; Professional, scientific and technical services

University:

University of Regina

Program: