Adversarial Machine Learning Engineer - Red Teaming

📋 Description

• Conduct hands-on adversarial testing across AI models, apps, agentic systems, and data pipelines to
• Perform red-team assessments: jailbreaks, guardrail bypass, prompt injection, tool-chain misuse
• Investigate edge-case findings and convert them into reproducible vulnerabilities.
• Assess risks like data poisoning, model inversion, membership inference, model extraction.
• Align security evals with OWASP Top10 for LLMs, NIST AI RMF, MITRE ATLAS, AI regs.
• Produce vulnerability reports with severity, evidence, steps, remediation tips.

🎯 Requirements

• Expert-level Python with PyTorch, TensorFlow, and Hugging Face Transformers.
• Hands-on fine-tuning of ML models/SLMs; LoRA, QLoRA, PEFT, instruction tuning, domain adaptation.
• Strong ML math: optimization, linear algebra, probability, statistics.
• Experience designing adversarial ML attacks: data poisoning, model extraction, membership inference.
• Experience building AI security defenses: adversarial training, robust fine-tuning, input
• Familiarity with Adversarial Robustness Toolbox, CleverHans, Foolbox.

🎁 Benefits

• Fully remote position available anywhere in Canada.
• Flexible environment focused on outcomes and impact.
• Work on advanced AI security challenges with foundation models.
• Clear career growth toward staff/principal-level influence.
• Supportive hiring process with guidance beyond.
• Collaborate with experienced AI security professionals.