At Snowflake, we are powering the era of the agentic enterprise. To usher in this new era, we seek AI-native thinkers across every function who are energized by the opportunity to reinvent how they work. You don’t just use tools; you possess an innate curiosity, treating AI as a high-trust collaborator that is core to how you solve problems and accelerate your impact. We look for low-ego individuals who thrive in dynamic and fast-moving environments and move with an experimental mindset — who rapidly test emerging capabilities to discover simpler, more powerful ways to deliver results. At Snowflake, your role isn’t just to execute a function, but to help redefine the future of how work gets done.

About the role

We are an AI-first analytics team. We don’t use AI to augment traditional BI workflows — we’ve replaced them. The Finance Analytics team builds the intelligence layer that Strategic Finance runs on: AI agents that encode repeatable finance processes, Streamlit apps that surface real-time insight, semantic models that let any analyst query complex data in plain English, and workflow automations that collapse hours of manual work into a single prompt.

Our primary development environment is CoCo (Cortex Code), Snowflake’s AI coding assistant, and SnowWork, the AI IDE we ship work in. Every deliverable on this team is built AI-first: you design the workflow, you write the prompt, you validate the output. If you are still building dashboards by hand, refreshing Excel files manually, or treating AI as a spell-checker for your code — this role will ask you to operate differently.

This is a high-breadth seat. One week you’re building a new AI agent for quarterly revenue analysis; the next you’re designing a sensitivity analysis tool for an earnings war room. You are equally comfortable in an AI-IDE, a Python file, and a stakeholder summary for a senior finance leader.

What you’ll work on

AI agent and workflow development (primary focus)

• Design and build skills and agentic experiences that encode repeatable finance workflows — revenue analysis, cost monitoring, earnings prep, headcount tracking — into reusable, invokable tools using CoCo and CoWork
• Write and iterate on prompt & skill structures (YAML + Markdown skill files) based on output quality and stakeholder feedback
• Build skills that allows non-technical finance analysts to produce analyst-quality output in a single prompt
• Evaluate model outputs rigorously — you are the quality gate before anything reaches a finance stakeholder

Finance analytics

• Build and maintain quarterly and weekly revenue summary pipelines
• Support sensitivity analysis models for quarterly business reviews & revenue forecast scenarios
• Produce ad-hoc analysis for Strategic Finance

Semantic Layer & Application development

• Own semantic layers end-to-end — model design, versioning strategy, verified query coverage, and accuracy iteration based on eval metrics; not just build models, but maintain the contract between the model and its consumers across each quarterly iteration
• Develop and deploy production finance dashboards as Streamlit apps (locally and deployed to Snowflake)
• Build customer-facing demo applications for Sales and Field teams
• Apply reusable component patterns and shared utility libraries for consistent, polished UI

Earnings and reporting automation

• Participate in quarterly earnings cycle prep — scenario tooling, export automation, IR data requests
• Build and maintain source-of-truth reporting exports (multi-tab Excel, formatted to spec)
• Support ad-hoc disclosure and investor relations data needs during quarter-end

Hard skills required

Must-have

AI-assisted development — You have used an LLM coding assistant (CoCo, Cursor, GitHub Copilot, Claude, or equivalent) as your primary development tool. You know how to write a prompt that produces production-ready output, how to steer a model that’s heading in the wrong direction, and how to encode domain logic into a reusable, parameterized skill. You have a measurable, trackable record of daily AI usage.

Prompt engineering and skill authoring — You can write a structured prompt (YAML + Markdown or equivalent) that routes correctly 95% of the time, handles edge cases gracefully, and encodes enough domain knowledge that the model behaves like a subject matter expert. You think in terms of context, instruction…