Principal Software Engineer - Snowflake Intelligence

Snowflake Snowflake · Data AI · CA-Menlo Park, United States · Engineering

Principal AI Engineer role focused on defining the technical vision and architectural design for Snowflake Intelligence, an AI product for enterprise data. The role involves leading the development of multi-agent systems, NL-to-SQL engines, and enterprise-scale RAG, with a strong emphasis on reliability, automated evaluation infrastructure, and cross-functional influence. It requires deep expertise in LLM orchestration, prompt optimization, guardrails, and integrating AI with traditional data stacks.

What you'd actually do

  1. Define the long-term technical vision for Snowflake Intelligence. Lead the architectural design of multi-agent systems, complex tool-use frameworks, and self-correcting NL-to-SQL engines.
  2. Move beyond simple evals to build world-class, automated "hill-climbing" infrastructure. You will establish the methodology for how Snowflake measures and guarantees LLM performance across diverse customer schemas.
  3. Partner with Product and Engineering leadership to align AI capabilities with business goals. You will bridge the gap between Research (modeling) and Production (infra), ensuring the latest LLM breakthroughs are viable at Snowflake scale.
  4. Design extensible "context engineering" patterns—including advanced function calling and semantic layer integration—that can be leveraged by other Snowflake teams and external developers.
  5. Act as a force multiplier. You will mentor Staff and Senior engineers, lead cross-functional "strike teams" on high-stakes projects, and foster a culture of technical excellence and rapid experimentation.

Skills

Required

  • 10+ years of software engineering experience
  • 3+ years specifically leading the deployment of LLM-based applications at massive scale
  • Deep, "under-the-hood" understanding of LLM orchestration
  • Expert in prompt optimization
  • Expert in semantic modeling
  • Expert in building robust guardrails for non-deterministic systems
  • Proven track record of designing complex systems that integrate AI with traditional data stacks (SQL, Retrieval Systems, Semantic Layers)
  • Exceptional ability to communicate complex technical trade-offs to both executive leadership and ICs
  • Held a "Principal" or equivalent high-level IC role, or have led the technical launch of a major AI product used by millions

Nice to have

  • Bachelor’s, Master’s, or PhD in Computer Science, AI, or a related field

What the JD emphasized

  • leading the deployment of LLM-based applications at massive scale
  • Deep, "under-the-hood" understanding of LLM orchestration
  • Proven track record of designing complex systems that integrate AI with traditional data stacks
  • Principal" or equivalent high-level IC role, or have led the technical launch of a major AI product used by millions

Other signals

  • agentic enterprise
  • agentic reasoning
  • NL-to-SQL
  • enterprise-scale RAG
  • LLM orchestration
  • multi-agent systems
  • tool-use frameworks
  • self-correcting NL-to-SQL engines
  • automated hill-climbing infrastructure
  • LLM performance across diverse customer schemas
  • context engineering
  • advanced function calling
  • semantic layer integration
  • deployment of LLM-based applications at massive scale