Post-Doctoral Fellow - Computer Science (Cybersecurity)

The Faculty of Business and Information Technology at Ontario Tech University has an opening for a Postdoctoral Fellow (PDF) in the area of Cybersecurity and AI. The position is offered for 12 months, preferably starting in Jan 1st, 2026 ‎subject to funding approval and may be extended contingent upon research performance.

This research program investigates the use of large language models (LLMs) with code-synthesis capability to automate metamorphic malware generation. The goal is to understand how modern foundation models - when fine-tuned on malware code corpora - can learn transformation rules that produce functionally-equivalent variants capable of bypassing static and signature-based detection.

The broader objective is to provide defensible, repeatable measurement of how AI-assisted metamorphic code generation may change the offensive-defensive balance - and to surface threat-relevant insights for defensive counter-design (e.g., more semantic detection and de-obfuscation strategies).

The position will be supervised by Dr. Pooria Madani.

Responsibilities/Accountabilities:

· Dataset preparation and domain-specific fine-tuning of LLMs for code mutation,

· Model compression/quantization to make such models small enough for endpoint-size deployment scenarios (threat modelling angle),

· Formal and empirical verification that mutated samples preserve semantics (correctness safety checks),

· Automated validation pipelines to measure detection evasion against commercial AV / EDR engines,

· Controlled experimentation to quantify detection degradation after each class of code-level metamorphic transformation.

Qualifications:

Required

· Strong experience training and fine-tuning large language models (LLMs), including prompt engineering and instruction tuning.

· Proficient with PyTorch and the Hugging Face ecosystem (Transformers, Datasets, Accelerate,Trainer or custom training loops).

· Solid software engineering skills for experiment implementation (Python, reproducible scripts, CI-friendly workflows).

· Practical knowledge of code-synthesis models and code representation/transformations.

· Deep understanding of malware concepts, especially metamorphic/obfuscation techniques and code-level mutation strategies.

· Experience with malware sandboxing and safe experimentation (dynamic analysis, instrumentation, VM/snapshot workflows).

· Strong research & technical writing skills (papers, reproducible experiment descriptions, responsible disclosure).

Preferred

· Experience with model compression/quantization (distillation, pruning, INT8/4 quant workflows).

· Familiarity with automated evaluation against AV/EDR products and red-team/blue-team methodologies.

· Background in formal verification or semantic-preserving program transformations.

· Prior publications or open-source contributions in ML for code, adversarial ML, or malware research.

· Awareness of ethics, legal, and responsible-research practices for dual-use work (IRB, disclosure, safe-lab protocols).