Machine Learning Security Researcher
About the Role
You will conduct cutting-edge security research on machine learning systems, identify novel attack vectors and failure modes, and develop tooling and evaluation frameworks to find and mitigate vulnerabilities. You will work with frontier model developers to assess robustness, build reproducible security tests and open-source tools, create threat models, publish findings, and present results to technical audiences. You will translate adversarial ML concepts into actionable recommendations and collaborate with engineering and research stakeholders to improve ML system security.
Requirements
- PhD or equivalent research experience in machine learning, deep learning, or related fields
- Expertise in adversarial machine learning and attack paradigms (evasion, poisoning, model inversion, membership inference, backdoors, prompt injection)
- Hands-on experience with ML frameworks such as PyTorch, JAX, and TensorFlow
- Deep technical knowledge of transformer architectures, training pipelines, and deployment/inference systems
- Familiarity with CUDA programming and GPU optimization (preferred)
- Strong software engineering skills in Python and at least one systems language (C/C++ or Rust)
- Proven research track record via publications, preprints, or open-source contributions
- Excellent communication skills and ability to present findings to technical and executive audiences
- Ability to rapidly learn new domains and think adversarially about complex systems
Responsibilities
- Conduct original security research on machine learning systems
- Identify novel attack vectors including adversarial examples, model poisoning, data extraction, and jailbreaks
- Assess the security posture of advanced ML systems for clients
- Design and build security testing frameworks, evaluation methodologies, and open-source ML security tools
- Develop comprehensive threat models and anticipate future attack vectors
- Publish research, present at conferences, and contribute to the research community
- Translate complex adversarial concepts into clear, actionable recommendations for stakeholders
Benefits
- Performance-based bonuses
- Fully company-paid insurance (health, dental, vision, disability, life)
- 401(k) plan with 5% employer match
- 20 days paid vacation
- 4 months parental leave
- $10,000 relocation assistance to New York City
- $1,000 working-from-home stipend
- $750 annual learning and development stipend
- Company-sponsored all-team celebrations including travel and accommodation
- Philanthropic contribution matching up to $2,000 annually