AI Engineer

AI Engineer - Hybrid This Jobot Job is hosted by:
Melanie Courtney Are you a fit? Easy Apply now by clicking the "Quick Apply" button and sending us your resume. Salary: $120,000 - $160,000 per year A bit about us:
Our client is a collection of industry-leading residential and commercial HVAC, electrical, and plumbing companies. Our mission is to provide exceptional residential and commercial services by upholding the legacies of our brand partners, empowering their growth, elevating performance, and enhancing the quality of life in the communities we serve. Why join us? Growing company Top benefits Great culture Job Details
1. Data Engineering & Snowflake Platform | Primary Focus Design and own the three-layer Snowflake architecture: RAW → STAGING → MART Build and maintain domain MARTs (operations, call center, finance, workforce, marketing) using dbt — grain, ownership, SLA, and data contracts defined per MART Own Snowflake RBAC — Domo read-only on MART, AI scoped to mart., row-level policies and masking applied where needed Build and maintain Semantic Views as the authoritative metric layer (booked rate, avg ticket, revenue per tech per day) — single source of truth for all tools Monitor pipeline health, warehouse costs, and query performance;

optimize proactively
2. Data Ingestion & ETL/ELT | Core Ownership Own ingestion end-to-end — Fivetran is the primary tool; configure and manage connectors for Service Titan, Quick Books/Sage Intacct, Paycom, CCaaS, and marketing platforms Handle sources Fivetran does not cover with lightweight Python scripts (SFTP, REST APIs, email-delivered files); bias toward replacing custom scripts with managed connectors over time Own the dbt project as an engineering artifact — structure, CI/CD, dev/prod environments, packages, documentation Manage schema drift — catch upstream source changes before they break downstream MARTs;

configure Fivetran schema change policies and dbt source assertions Define and maintain pipeline SLAs and alerting — failures surface before business users notice Own data lineage — every MART table traceable to source via dbt lineage graph Manage Fivetran MAR costs — optimize sync frequency without sacrificing freshness 3. AI Engineering | Built on the Data Foundation Implement Cortex Analyst + Semantic Views for NL-to-SQL querying of MARTs — no SQL required for GMs and ops users Build enterprise knowledge search via Cortex Search — SOPs, runbooks, and operational docs indexed within the Snowflake perimeter Deploy Cortex Agents orchestrating across Cortex Search and Cortex Analyst for multi-step business questions Connect external systems (Service Titan, M365) via Snowflake MCP Server;

use N8N or Lang Chain for workflows outside Snowflake Build AI feature pipelines in mart. under the same governance standards as all other MARTs
4. Governance & Enablement Define AI governance framework: acceptable use, risk tiers, PII handling, vendor evaluation Ensure AI pipelines inherit existing Snowflake RBAC and masking — no parallel governance layer Train end users on CoWork and Cortex Analyst; build internal tools for ops, finance, and service teams First 90 Days Audit Domo pipelines and produce a source inventory — ingestion method, refresh schedule, and downstream consumers for every source Stand up Snowflake three-layer architecture, RBAC, and dbt project scaffold with dev/prod environments Configure first Fivetran connectors (Service Titan first);

ship mart.operations with dbt tests passing and Domo read-only connected Deliver AI governance framework and acceptable use policy Present 12-month roadmap to leadership — ingestion milestones gating MART build gating AI capabilities Our Tech Stack Data Warehouse (required) Snowflake + Cortex AI (Cortex Agents, Cortex Search, Cortex Analyst, Semantic Views) Data Ingestion (required) Fivetran (strongly preferred);
Airbyte or equivalent considered;
Domo Writeback (transitional);
Python for uncovered sources Data Transformation (required) dbt — models, tests, sources, snapshots, incremental, CI/CD BI / Visualization Domo (read-only MART layer) AI / LLM Claude (Anthropic) — direct API and via Cortex…