Staff Research Engineer (LLM Pre-Training)

📋 Description

• Train foundation models for coding tasks.
• Develop LLMs from scratch and deploy to production.
• Make models accessible to end users worldwide.
• Convert business requirements into technical specs.
• Train LLMs on a large GPU cluster.
• Collect and process pre-training and fine-tuning data.

🎯 Requirements

• Production ML systems design, deployment, and support.
• Strong NLP theory and transformer-based approaches.
• Proficiency with PyTorch and NLP libraries.
• Distributed training of multi-billion parameter models.
• Experience with LLM inference frameworks (vLLM, DeepSpeed, TensorRT).
• RLHF/RLAIF alignment techniques.