Forward Deployed Analytics Engineer

Improvado•,

18d•$70,000 - $120,000•Remote

About The Position

Improvado is an AI-powered marketing data platform built for mid-market and enterprise teams. We help companies automate complex marketing data workflows, unify data at scale, and surface insights through BI and AI. Brands like ASUS, Docker, Activision, and H&R Block rely on us to simplify analytics and improve marketing performance. We're a fast-growing Series A startup backed by $34M in funding and driven by a team that moves fast, stays curious, and cares about high-quality execution. You join a customer's marketing data problem and stay until it runs in production. One month it's a pharma brand that needs validated files delivered to its agency every night. The next it's a DTC brand that wants attribution it can defend, or an agency running forty client dashboards on one model. You own three to six accounts end to end: the discovery calls, the data model, the pipeline, the dashboard, the handoff. Claude Code is your pair, not a checkbox.

Requirements

3+ years shipping data work for people outside your own team: an agency, a consultancy, professional services, or an in-house analytics team with internal customers.
SQL is your first language and Python a comfortable second.
You've used Claude Code, Codex or Cursor for real work and can say where they go wrong.
You can hold a call with a marketing director and a data engineer in the same hour and both leave with a decision.
English C1 or better.

Nice To Haves

Ad-platform, GA4 or CRM data.
dbt in production.
MMM or attribution.
Temporal or Airflow.

Responsibilities

Run weekly customer calls and turn what you hear into a scoped build with dates.
Design and maintain the models: SQL and dbt on ClickHouse, BigQuery or Snowflake, kept correct as sources change.
Write Python pipelines on Temporal when the platform doesn't cover the case: SFTP feeds, validation rules, custom joins, audit emails.
Ship dashboards in Looker, Power BI or Tableau, whichever the customer already runs.
QA your own delivery before the customer sees it.
Send platform gaps back to Product with a reproducible case.