Snowflake is hiring a Staff Research Scientist for AI Agents and LLMs, US remote. You serve as a technical pillar in Snowflake's AI research org. You don't just execute on roadmaps, you define them. Work sits at the intersection of autonomous agents and LLMs, building the models and systems that turn enterprise data into self-directed, decision-making agents. The team has delivered Arctic LLM (enterprise foundation models), Arctic-Text2SQL (frontier-quality reasoning at low cost/latency), Snowflake Intelligence (multi-step reasoning + Deep Research), Arctic Inference (open-sourced LLM serving stack), Arctic Long-Context Training (1M+ token training on a single H200), and Agent World Models (RL training across 1000+ synthetic environments). You: define research direction in Agentic AI and LLMs, train and fine-tune models for enterprise reasoning and tool use, advance state-of-the-art capabilities in multi-step reasoning, advance retrieval/grounding/memory/multi-agent coordination, establish evaluation standards. Required: deep ML research background, hands-on LLM training, publication track record. What stands out: real research output that ships as products (Arctic LLM, Snowflake Intelligence). Not academic-only.