for Teaching Assistants, Graduates Service Assistants & Sessional Lecturers

SL402570-Winter 2027-CIS*6670 ST: AI and LLM Security and Privacy

Sessional Lecturer, Unit 2

Academic Unit:

School of Computer Science

Semester(s) of Assignment(s):

Winter 2027

Number of Available Work Assignment(s) / Sections:

1

Level of Work Assignment(s):

1

Right of First Refusal (RoFR)

A Sessional Lecturer holds a RoFR (i.e., for a particular course) if they have successfully taught the course in the past six (6) semesters. A SL who holds a RoFR to this course is required to exercise their right by way of the online hiring system. Also see: What is Right of First Refusal (RoFR)?

A Sessional Lecturer Currently Holds a Right of First Refusal for this Course:

No

Course Number:

CIS*6670

Course Name:

ST: AI and LLM Security and Privacy

Course Format:

In-Person

Course Description:

See Course Calendar

Other Course Description or Assignment Information:

- Large language models (LLMs) and other AI systems are now deployed in critical areas such as health, finance and cybersecurity, and they introduce recent security and privacy risks. This offering of Special Topics examines attacks on machine learning and LLM systems, the defences against them, and the governance needed to deploy them safely. It moves from classical adversarial machine learning to LLM-specific risks, then to privacy, provenance and governance.Topics include:- threat modelling for ML and LLM systems (NIST AI 100-2 taxonomy);- evasion, poisoning and backdoor attacks;- privacy attacks,



including membership inference and training-data extraction;- jailbreaking and safety alignment;- direct and indirect prompt injection and their defences;- poisoning of retrieval-augmented generation (RAG) and the model supply chain;- security and privacy of LLM agents;- privacy-preserving machine learning (differential privacy, DP-SGD, privacy auditing, federated learning);- machine unlearning, watermarking and provenance of AI-generated content;- LLMs for security and their dual-use risks;- AI governance and risk management (NIST AI RMF, EU AI Act, OWASP Top 10 for LLM Applications and the Canadian regulatory context).

60

Anticipated Duties and Responsibilities

- Orientation-Training
- Office Hours
- Preparation
- Student Consultation
- Conducting Labs/Seminars
- TA Coordination Meetings
- Invigilating Exams
- Grading

Qualifications

Required Qualifications

Degree:

PhD related to field

Other

Masters in Computer Science/Engineering or related field

Prior Teaching Experience:

Successful teaching related to field at college or university level

Required competence, capability, skill and ability related to course content:

Research and practical experience in AI and LLM security and privacy,



including experience with:- threat modelling of ML/LLM systems;- adversarial machine learning (evasion attacks, data poisoning, backdoors);- privacy attacks (membership inference, training-data extraction);- LLM jailbreaking and prompt injection (direct and indirect);- RAG systems and LLM agent security;- differential privacy and DP-SGD, privacy auditing and federated learning;- machine unlearning and watermarking of LLM outputs;- deep learning frameworks and model ecosystems (Python, PyTorch, Hugging Face);- deploying and evaluating open-weight LLMs;- AI risk and governance frameworks (NIST AI RMF, NIST AI 100-2, EU AI Act, OWASP Top 10 for LLM Applications).

Preferred Qualifications

Degree:

PhD and expert in course content

Other

PhD in Computer Science, Computer Engineering or a related field

Prior Teaching Experience:

Many years of successful teaching related to contents of the course.

Other

Successful teaching of graduate courses in AI security, machine learning security or privacy

Specific Preferred competence, capability, skill and ability related to course content:

- Experience with LLM red-teaming and evaluation tools (e.g., garak, PyRIT, AgentDojo, HarmBench) and privacy-preserving ML libraries (e.g., Opacus);
- Experience applying LLMs to security operations (e.g., SOC analytics, vulnerability analysis);
- Industry or applied experience securing deployed AI systems will be considered an asset.

Days Required and Wages

Days and Times Required:

Dates and times have not been finalized, bu

📌 ST: AI and LLM Security and Privacy (Canada)
🏢 University of Guelph
📍 Canada