🛠 Tech Stack

💼 About This Role

You'll develop training optimization techniques for large-scale models, improving efficiency and stability. You'll design experiments and publish research at top venues. This role offers direct influence over core model training decisions.

🎯 What You'll Do

• Design and evaluate training optimization techniques for large models
• Improve training efficiency and stability across long runs
• Run large-scale experiments and analyze results
• Author research papers and collaborate with infrastructure teams

📋 Requirements

• Strong background in machine learning research with emphasis on training dynamics
• Experience training large neural networks (LLMs, multimodal models)
• Publication experience in ML venues (NeurIPS, ICML, ICLR, etc.)
• Proficiency in Python and PyTorch

✨ Nice to Have

• Experience with non-standard architectures (RNN variants, long-context models)
• Experience optimizing training on GPUs at scale (FSDP, ZeRO, custom kernels)
• Contributions to open-source ML or research codebases

🎁 Benefits & Perks

• 🚀 Real influence over core model training decisions
• 📄 Publish novel research with freedom to pursue ideas
• 💻 Direct access to large-scale experiments and production constraints
• 👥 Small, senior team that values deep thinking