ML Privacy Engineer — LLM Safety & Data Security

Capitolis; San Francisco (CA); On-site; USD 150,000 - 210,000; Full time

Job summary

Dynamo AI is seeking a highly capable ML privacy specialist to own privacy-focused initiatives, from detection of data leakage to PII protection, and to apply algorithms in customer-facing applications with engineering collaboration.

You will generate synthetic data, train LLMs, and rigorously evaluate and benchmark models, contributing to responsible AI deployment and privacy-preserving technology leadership.

Qualifications

• Deep domain knowledge in privacy-preserving ML.
• Practical experience in techniques to attack or defend ML models in terms of privacy.
• Extensive experience in implementing multiple different types of LLM models and architectures in the real world.
• Adaptability and flexibility in switching focus with new findings in the community.
• Preferred: previous projects or research in LLM privacy.

Responsibilities

• Own an ML privacy vertical e.g. data leakage attacks, sensitive PII detection, and/or membership inference attacks.
• Collaborate with our engineering team to deliver real-world applications of your algorithms for our customers.
• Generate high quality synthetic training data, train LLMs, and conduct rigorous evaluation and benchmarking.

Skills

Privacy-preserving ML

ML privacy attacks/defense

LLM architectures

Adaptability

LLM privacy background

Job description

Dynamo AI is seeking a highly capable ML privacy specialist to own privacy-focused initiatives, from detection of data leakage to PII protection, and to apply algorithms in customer-facing applications with engineering collaboration.

You will generate synthetic data, train LLMs, and rigorously evaluate and benchmark models, contributing to responsible AI deployment and privacy-preserving technology leadership.