Senior Machine Learning Engineer, AI Safety
Core
Scaling safety for multi-modal Large Language Models (LLMs) including advanced agentic safety, focusing on LLM Security, Frontier Risks, and Multi-turn Safety Evaluation.
Role type
Senior Machine Learning Engineer (AI Safety)
Builds
Production-ready safety datasets, training recipes, and evaluation algorithms for frontier models.
Domain
Artificial Intelligence / Large Language Models / AI Safety
Deliverable
production ML models
Required skills
Post-training of LLMs (SFT, RLHF/RLAIF), LLM Security (backdoors, poisoning), Frontier Risks (deception, manipulation), Agentic Safety, Multi-turn Safety Evaluation, PyTorch, Python, Multimodal Systems
Preferred skills
Published papers on AI Safety at top-tier conferences, Open-source contributions to AI Safety tools, Alignment of Vision-Language Models (VLMs)
Responsibilities
Develop datasets and specialized models to evaluate/benchmark models in safety tracks; Develop datasets and recipes for filtering training data and RL environments; Research and deploy new approaches like Instruction Hierarchy or Risk Detection; Partner with cross-functional teams to scale safety solutions.
Seniority
Senior, hands-on IC
