🛠 Tech Stack

💼 About This Role

You'll work on data generation and post-training algorithms for safety in LLMs that access external resources. You'll improve the trustworthiness and security of next-gen models.

🎯 What You'll Do

• Design and run data collection tasks with human annotators.
• Develop evaluation methods for model safety and robustness.
• Implement post-training algorithms to improve model behavior.
• Analyze dataset quality, biases, and training suitability.

📋 Requirements

• Strong statistical skills for evaluating scientific experiments.
• Strong software engineering skills with Python and ML frameworks.
• Experience designing data collection tasks with human annotators.
• Hands-on experience training LLMs on distributed infrastructure.

✨ Nice to Have

• Familiarity with evaluating ML system generalizability.
• Papers at top-tier venues (NeurIPS, ICML, etc.).

🎁 Benefits & Perks

• 🤝 Open culture and inclusive work environment.
• 🧑‍💻 Cutting-edge AI research team.
• 🍽 Weekly lunch stipend and in-office lunches.
• 🦷 Full health and dental with mental health budget.
• ✈️ 6 weeks vacation (30 working days).