Adversarial Machine Learning Engineer - Red Teaming
J
JobgetherAI Security
CanadaFull-Time
Salary not disclosed
Apply NowOpens the employer's application page
Job Details
- Required Skills
- PythonMachine LearningPyTorchTensorflowLLMMLOps
Requirements
- Expert-level Python programming skills with strong experience using machine learning frameworks such as PyTorch, TensorFlow, and Hugging Face Transformers.
- Hands-on experience fine-tuning machine learning models and Small Language Models (SLMs), including techniques such as LoRA, QLoRA, PEFT, instruction tuning, and domain adaptation.
- Strong foundation in machine learning mathematics, including optimization, linear algebra, probability, and statistics.
- Proven experience designing and executing adversarial machine learning attacks, including adversarial examples, data poisoning, model extraction, and membership inference.
- Experience implementing AI security defenses such as adversarial training, robust fine-tuning, input sanitization, and differential privacy.
- Familiarity with adversarial machine learning frameworks such as Adversarial Robustness Toolbox, CleverHans, and Foolbox.
- Experience conducting AI and LLM red-team exercises, including jailbreak testing, prompt injection assessments, and safety evaluation.
- Ability to evaluate and benchmark AI model robustness, security posture, and safety before and after fine-tuning.
- Experience with MLOps practices, including model versioning, experiment tracking, and secure model deployment pipelines.
- Strong threat-modeling skills with the ability to think like an attacker while communicating risks effectively.
- Ability to collaborate with technical and non-technical stakeholders in a fast-moving environment.
Responsibilities
- Conduct hands-on adversarial testing across AI models, applications, agentic systems, and data pipelines to identify security vulnerabilities and weaknesses.
- Perform advanced red-team assessments including multi-turn jailbreaks, guardrail bypass testing, prompt injection analysis, agent and tool-chain misuse evaluation, and dangerous capability assessments.
- Investigate edge-case findings from AI security campaigns and transform anomalies into fully understood, reproducible vulnerabilities.
- Assess risks related to data poisoning, model inversion, membership inference, model extraction, and other adversarial machine learning threats.
- Develop and execute security evaluations aligned with industry frameworks such as OWASP Top 10 for LLM Applications, NIST AI Risk Management Framework, MITRE ATLAS, and relevant AI regulations.
- Produce detailed vulnerability reports with severity ratings, evidence, reproduction steps, and actionable remediation recommendations.
- Partner closely with AI safety, engineering, and security teams throughout vulnerability discovery, remediation, and validation processes.
- Retest implemented fixes to confirm that security improvements are effective and vulnerabilities have been properly addressed.
- Communicate technical findings clearly to both engineering teams and non-technical stakeholders by translating complex AI risks into understandable business impacts.
View Full Description & ApplyYou'll be redirected to the employer's site
