Applied Research Engineer - Post-training

The role is for a Research Engineer in post-training at Magic, focusing on developing novel techniques and datasets to improve model performance for real-world applications.

• Research and develop post-training techniques
• Enhance model capabilities for code generation
• Ensure seamless integration for users
• Create continuous data feedback mechanisms
• Design scalable approaches for synthetic datasets

• Experience with LLM productization
• Strong software engineering skills
• Deep learning literature knowledge
• Post-training experience with LLMs
• Ability to generate and evaluate research ideas
• Attention to detail and data quality
• Proficiency in debugging large ML codebases