In-context Learning of Software Vulnerability Detection Using LLMs

Future / MS research topic exploring in-context learning, few-shot prompting strategies, and domain adapter tuning for software vulnerability detection across multi-language code bases.

In-context LearningVulnerability DetectionLLM Benchmarking

Availability

Open for BS SupervisionCOMSATS University Islamabad, Lahore

Supervision

Dr. Junaid AkramLead Supervisor

Institution

Vulnerability DetectionCOMSATS University Islamabad, Lahore Campus

Supervision prerequisites

Recommended Prerequisites

Strong programming and software engineering skills.

PythonHugging FaceCodeBERT / StarCoderCVE Corpus

Source & Verification

Official Listing Details

Source: Official COMSATS Supervision Listing

Last Verified: July 2026

Request Supervision

Supervised Projects Portfolio

More Supervised Projects

View all projects

Full-Stack & Systems

MegiLance — AI and Blockchain-Powered Freelancing Platform

A supervised final-year engineering development project combining a full-stack freelancing marketplace with AI decision assistance and blockchain payment escrow.

Ghulam Mujtaba, Muhammad Waqar Ul MulkView Case Study

Vulnerability & Security

Software Vulnerability Detection using LLM Model

A running research project leveraging Large Language Models (LLMs) to detect complex software vulnerabilities in source code through fine-tuning and AST-based prompt engineering.

Syed Muhammad Irtaza Shahid, Muhammad Waleed, Muhammad Muneeb AzharView Case Study

Vulnerability & Security

Token-Based Semantic Code Clone Detection

A running software security project building a scalable token-based semantic code clone detection platform to capture Type-3 and Type-4 clones across large open-source repositories.

Ali Muhammad, Ali Murtaza, Hafiz Muhammad Shams ul HudaView Case Study