- Should be able to fine-tune models, especially in contexts where the architecture differs from standard setups. 1
- Must guide the team in reinforcement learning and resolve issues during fine-tuning, articulating solutions clearly. 2
- Needs experience in deep learning training; prior LLM fine-tuning is a plus but not mandatory, as deep learning experience is considered sufficient to understand LLM fine-tuning requirements. 3
- Should have expertise in advanced agentic AI, including multi-step RAG (Retrieval-Augmented Generation) solutions, architecting tools, MCP servers, writing advanced RAG code, and benchmarking both fine-tuned and RAG models. 4
Senior AI Engineer Expectations:
- Should consult with the principal engineer to develop advanced RAG models, not just basic or naive RAG implementations. 1
- Expected to build multi-step or advanced RAG systems, including features like self-reflection, grading, and using cross-encoders or bi-encoders. 2
- Some engineers should be skilled in RAG development, while others should focus on fine-tuning models, with a mix of expertise across the team