🛠 Tech Stack

💼 About This Role

You'll lead high-impact research on AI safety at the Center for AI Safety (CAIS), tackling ambitious open problems and driving them to publication. You'll design experiments on large language models, build tooling for training and evaluation, and collaborate with external partners using our compute cluster to reduce real-world risks from advanced AI.

🎯 What You'll Do

• Pursue high-value research directions with limited guidance.
• Own end-to-end research experiments from conception to publication.
• Train and fine-tune large transformer models across domains.
• Build and maintain datasets, benchmarks, and distributed training pipelines.

📋 Requirements

• PhD in computer science, ML, or related field
• 5+ years of related research experience
• Extensive experience with PyTorch and Hugging Face
• Strong publication record at top venues (NeurIPS, ICML, ICLR, etc.)

✨ Nice to Have

• Experience with distributed ML job launching and training
• Familiarity with AI safety subfields like honesty, robustness, or transparency

🎁 Benefits & Perks

• 🔬 Cutting-edge research at a top AI safety organization
• 🖥️ Access to large compute cluster for experiments
• 🤝 Collaboration with academic and commercial partners
• 📝 Publication support at top conferences