Senior Data Intelligence Engineer

Deepgram

About The Position

Deepgram is the foundational AI company for voice. We build the models that allow machines to hear, understand, and speak to humans with zero latency. As we scale our usage-based economy, we are building the "Intelligence Layer"—a system of autonomous agents that monitor, reason, and act on our data to drive NRR and operational excellence. The Role Reporting to the Head of Decision Intelligence, you are the founding engineer of the Data Intelligence team. While the Head of Data sets the strategic "Brain" of the company, you build the nervous system. You are not building legacy dashboards; instead you are building the API-based tools, dbt models, and semantic layers that allow AI agents to navigate our data environment autonomously.

Requirements

Experience: 5+ years in high-growth Data or AI Engineering roles.
AI-Native Technical Skills: Mastery of SQL and Python (production-grade) with a proven track record of building RAG or Agentic systems on top of structured data.
Usage-Based SaaS Fluency: Familiarity with the metrics of an API-first business, such as consumption patterns, gross margins, and NRR.
Builder Mindset: Comfort with "work-in-progress" infrastructure and the ability to turn messy, siloed data into high-impact business logic.

Nice To Haves

Prior experience as a Software Engineer or Data Engineer.
Experience with real-time billing and consumption systems.
Contributions to open-source data or AI projects.

Responsibilities

Semantic Layer Architecture: Build and maintain high-fidelity dbt and SQL models that serve as the "ground truth" for our complex, usage-based revenue models.
Agent Tooling & Integration: Develop the tools and permissions frameworks that allow "Analyst Agents" to query Athena, correlate Salesforce churn signals, and identify API latency issues.
Infrastructure Evolution: Serve as the technical bridge to the Engineering/Infra team to ensure our data contracts are "agent-ready" and highly reliable.
The "Unlock Audio Intelligence" Project: Partner with the Head of Data to ingest thousands of hours of internal calls using Deepgram’s own models, turning unstructured audio into queryable insights for GTM teams.
Automation Obsession: Maintain a culture where manual, repetitive SQL tasks are viewed as automation bugs to be solved with code and agents.