🛠 Tech Stack

💼 About This Role

You'll advance post-training of large language models using SFT, RLHF, and continual pretraining to align models for reliable JSON-schema function calls. You'll also design and deploy MCP servers for checkpoint routing and safety.

🎯 What You'll Do

• Advanced post-training of large language models (SFT, RLHF, continual pretraining).
• Aligning models for reliable JSON-schema function calls and tool usage.
• Design, deploy, and operate Model Context Protocol (MCP) servers.
• Build offline and live eval pipelines for alignment and factuality.

📋 Requirements

• 3+ years of experience in LLM development and optimization.
• Proven track record in advanced post-training techniques (SFT, RLHF, RLAIF).
• Hands-on experience with distributed training frameworks (DeepSpeed, FSDP).
• Proficiency in Python and machine learning frameworks (PyTorch, TensorFlow).

✨ Nice to Have

• Experience with model deployment on vLLM or Triton clusters.
• Knowledge of JSON-schema and API development.