Principal AI Platform Engineer, Applied AI

Snowplow•,

About The Position

Snowplow is hiring a Principal AI Platform Engineer for its Applied AI team, which owns Firn, the company-wide data and AI platform. Firn provides governed access to Snowflake through an MCP gateway, routes model calls through an LLM gateway with tracing, supports agentic workflows through a factory framework and run ledger, and maintains the company’s skills catalog. The platform is at an early stage, with plans to expand it to support commercial intelligence and safe agent actions. The role leads the expansion of Firn, beginning with a full-time engagement focused on Go-to-Market for about six months. The engineer will bring existing GTM services onto the platform, unify and modernize them, build agents, tools, and learning loops, and make the services production-grade. The work includes designing platform services with data engineers and supporting internal customers across Marketing, Sales, Account Management, and Partnerships. The Principal Engineer reports to the Senior Director of Applied AI and works as a peer. Internal customer teams set their priorities, while the engineer determines how the platform is built and ensures its technical soundness. After the GTM services run independently, the engineer remains their owner while focusing primarily on the company’s next platform needs.

Requirements

Extensive production software experience, including designing, shipping, and operating systems that other teams depend on; Python or TypeScript, SQL, CI/CD, and cloud infrastructure used as everyday tools.
Experience building production LLM systems, including agents with tools, MCP servers, gateways, tracing, and evaluation, with the ability to discuss production failures and changes made in response.
Experience building platforms used by other teams or customers through designed interfaces, and the ability to distinguish a platform from a collection of integrations.
Fluency with data platforms, including Snowflake or an equivalent warehouse, dbt, data contracts, and pipelines moving warehouse data into and out of SaaS tools.
Operational discipline with alerting, recovery, runbooks, and cost tracking, including experience operating systems where these practices mattered.
Ability to judge which decisions should be handled by deterministic code, an agent, or a person, and to explain the reasoning to non-engineers.
An AI-native workflow, agent, or tool that the candidate built and still uses; the candidate should expect to demonstrate it during the interview process.
Ability to work with non-engineer internal customers, gather requirements, and maintain sound technical implementation choices; sales or marketing background is not required.
Experience setting technical direction for a platform that other engineers followed based on the quality of the design rather than reporting relationships.

Responsibilities

Design and own the GTM intelligence engine on Firn, including the data and AI used to determine whom to engage, how to engage them, and how to learn from outcomes.
Work with data engineers to design GTM capabilities as platform services for multiple internal customers and bring existing GTM services onto the shared foundation.
Build agents that run GTM work and tools that people and agents can use, using the factory framework with review steps and evaluations as standard practices.
Own and evolve platform architecture, including the MCP gateway, LLM gateway, tracing, factory framework, and ledger; decide what to build, buy, or retire.
Establish production-grade practices for platform services, including version control, CI, alerting, recovery, cost visibility, documentation, and handover.
Move rules from individual tools into versioned platform code that agents can read and engineers can test, turning tools into endpoints driven by the platform.
Ensure platform actions are recorded and tied to outcomes to enable learning loops; own instrumentation that applies metric definitions established with data engineers.
Own evaluation and tracing tooling used by embedded and forward deployed engineers to validate their agents.
Set the engineering bar for the Applied AI team, own the platform roadmap, and maintain a long-term view of the platform.
Remain the owner of GTM services after they are running independently, while directing most subsequent effort toward the company’s next needs.