🛠 Tech Stack

💼 About This Role

You'll develop and tune post-training recipes for large models at Thinking Machines Lab, a team behind AI products like ChatGPT and PyTorch. This role sits at the core of the roadmap, blending research and engineering to make AI collaborative and safe. You'll iterate on training recipes, evals, and scaling methodologies while publishing impactful research.

🎯 What You'll Do

• Develop and tune post-training recipes with datasets and hyperparameters.
• Iterate on evaluations to ensure metrics capture what matters.
• Debug training configurations and analyze unexpected results.
• Scale existing methodologies and explore new training approaches.

📋 Requirements

• Proficiency in Python and a deep learning framework like PyTorch, TensorFlow, or JAX.
• Comfortable debugging distributed training and writing scalable code.
• Bachelor's degree in CS, ML, Physics, Math, or related field with strong theoretical grounding.
• Clear communication of complex technical concepts in writing.

✨ Nice to Have

• Strong grasp of probability, statistics, and ML fundamentals.
• Experience with RLHF, RLAIF, or preference modeling for large models.
• PhD in relevant field or equivalent industry research experience.

🎁 Benefits & Perks

• 🏖️ Unlimited PTO
• 🤒 Health, dental, and vision benefits
• 👶 Paid parental leave
• 🚚 Relocation support
• 💰 Visa sponsorship

📨 Hiring Process

We continuously review applications and reach out to applicants as new opportunities open; please avoid applying more than once every 6 months.

🚩 Heads Up

• Evergreen role with no immediate position may lead to long wait.
• No specific years of experience given, making level unclear.