Machine Learning Auditor

Dijav AI Group; Dover (DE); On-site; USD 120,000 - 180,000; Full time

Job summary

DijavAI is seeking a specialized Machine Learning Auditor to perform deep-dive technical audits of complex neural networks. You will assess explainability, robustness, and security vulnerabilities across ML pipelines.

You will conduct code reviews, test models against adversarial attacks, validate explainability methods like SHAP/LIME, verify data lineage, and automate audit tasks using Python. This role requires communicating with non-technical stakeholders.

Qualifications

• PhD or Master’s in CS/AI/ML required or strongly preferred.
• Experience auditing ML models and pipelines.
• Ability to explain complex concepts to non-technical stakeholders.

Responsibilities

• Perform code reviews and architecture analysis of ML pipelines.
• Test models for adversarial attacks and security vulnerabilities.
• Validate model explainability methods (SHAP, LIME, etc.).
• Verify data lineage and provenance tracking mechanisms.
• Automate parts of the technical audit process using Python scripts.

Skills

Code reviews

Explainability

Security auditing

ML pipelines

Education

PhD or MS in CS/AI/ML

Tools

PyTorch

TensorFlow

MLops tools

Job description

We are looking for a specialized Machine Learning Auditor to perform deep-dive technical audits of complex neural networks. You will focus on explainability, robustness, and security vulnerabilities.

What You'll Do

• Perform code reviews and architecture analysis of ML pipelines.
• Test models for adversarial attacks and security vulnerabilities.
• Validate model explainability methods (SHAP, LIME, etc.).
• Verify data lineage and provenance tracking mechanisms.
• Automate parts of the technical audit process using Python scripts.

What We're Looking For

• PhD or strong Master's in Computer Science or AI/ML.
• Experience with PyTorch, TensorFlow, and MLops tools.
• Background in cybersecurity or algorithmic auditing.
• Ability to explain complex technical concepts to non-technical stakeholders.
• Published research in AI safety or fairness is highly desirable.