Data And Analytics Jobs in San Francisco, CA

About Broccoli

Broccoli AI is building the AI operating system for home service businesses.

We work with plumbing, HVAC, roofing, etc. contractors, the people who keep homes running, and replace fragmented tools and manual workflows with AI agents that actually do the work.

Our AI assistants answer phones, book jobs, follow up with customers, and drive revenue, fully integrated into systems like ServiceTitan.

We started by going door-to-door, meeting 100+ contractors, and understanding how these businesses actually run. That shaped everything we've built. Today:

• Hundreds of contractors use Broccoli to run their front office.

• We've grown from $0 to millions in ARR in under a year.

• We're trusted by both single-location operators and the largest PE-backed roll-ups.

• We've raised $25M+ from Khosla Ventures and Y Combinator.

About the role

Our data is rich and comes from many sources. Every customer dashboard, every business review, every metric the company runs on draws from it — and we haven't yet built the unified data layer that makes all of that fast, consistent, and ready to scale.

You'll build that layer and own it. You'll model our data into clean, documented tables and build the source-of-truth library and own the data definitions the whole team runs on. You'll work hand in hand with the Strategy & Ops team — and essentially every tool we build, especially the external-facing dashboards and analytics our customers see, will be built on your work.

You're the team's first dedicated data hire: you own the architecture, the tooling choices, and the trust in every number. What you build powers customer-facing dashboards, cross-customer benchmarks, and eventually the business intelligence we ship inside the product.

What you'll do

• Build and run the pipelines. Reliable ingestion from all our sources into ClickHouse — you choose the tooling and own the flow.

• Model the data. Turn raw feeds into clean, documented tables — including entity resolution, so a customer is the same customer across billing, support, and call data.

• Build the source-of-truth library. Canonical views and metric definitions that every dashboard and analysis reads from.

• Make the data Human & AI-ready. Structure our models, definitions, and documentation so both people and AI agents can query them and get the right answer — then build the internal tools that let anyone at Broccoli ask a data question and trust the response.

• Keep it trustworthy. Freshness checks, quality tests, and alerts — we find out a pipeline broke before a customer does.

• Run deep dives when the team needs them. Ad-hoc analyses, segment investigations, partner questions.

• Work closely with engineering. Understand how our systems store and produce data, including schemas, events, and architecture, and give input early on changes so the data that lands in the warehouse is usable, stable, and easy to model.

What we're looking for

• 4–8+ years in data or analytics engineering — you've built and operated production pipelines end to end, and been the one paged when they broke.

• Strong SQL and solid Python; hands-on with ETL tooling (Airbyte, Fivetran, Dagster, dbt, or hand-rolled) and orchestration.

• Real experience with a columnar/OLAP warehouse — ClickHouse ideally; BigQuery, Snowflake, or Redshift transfer fine.

• Data modeling as a craft: you've designed the tables other people query, and you care what the numbers mean, not just that the pipes run.

Nice to have

• Self-directed: you've been the first or only data person somewhere, or built a data platform from scratch

• ClickHouse specifically — materialized views, performance tuning on event-scale data.

• Multi-source identity / entity resolution experience.

• Exposure to customer-facing or multi-tenant analytics (strict customer-level data isolation).

• B2B SaaS operational data — calls, bookings, jobs, billing — or CRM/field-service data like ServiceTitan.