🛠 Tech Stack

💼 About This Role

You'll own the red-teaming and adversarial evaluation pipeline for Reflection's open-weight AI models, probing for failure modes across security, misuse, and alignment gaps. You'll work hand-in-hand with the Alignment team to translate safety findings into concrete guardrails for every release. This role is a critical gatekeeper for open-weight model shipping.

🎯 What You'll Do

• Own the red-teaming and adversarial evaluation pipeline for models.
• Translate safety findings into concrete guardrails with Alignment team.
• Validate that every release meets risk thresholds before shipping.
• Develop scalable, automated safety benchmarks and adversarial tests.

📋 Requirements

• Graduate degree (MS/PhD) in CS, ML, or equivalent in AI Safety.
• Deep technical understanding of LLM safety and adversarial attacks.
• Strong software engineering for automated evaluation pipelines.
• Experience with RLHF/RLAIF preferred.

✨ Nice to Have

• Experience with reinforcement learning (RLHF/RLAIF).
• Experience building large-scale ML systems.

🎁 Benefits & Perks

• 💰 Top-tier compensation with salary and equity.
• 🏥 Comprehensive health & wellness including medical, dental, vision.
• 👶 Fully paid parental leave and family planning support.
• 🏖️ Paid time off and relocation support.
• 🍽️ Daily lunch and dinner provided, plus off-sites.