Forward Deployed Marketing Data Scientist

Hightouch Hightouch · Data AI · San Francisco, CA · Engineering

Role focuses on partnering with customers and internal engineering teams to ensure AI-driven marketing campaigns deliver measurable impact. This involves diagnosing model behavior, tuning ML levers, analyzing incrementality, exploring customer data, and explaining insights to marketers and executives. The core task is to ensure AI agents perform optimally and to help customers understand their performance.

What you'd actually do

  1. Explain _why_ AI Decisioning is driving lift using counterfactuals, incrementality breakdowns, and cohort analysis.
  2. Debug performance issues, iterate on reward functions, and ensure the agent’s recommendations align with customer goals.
  3. Pull down historical data to run exploratory analyses using Polars / Pandas in Jupyter notebooks
  4. Modify and improve customer feature matrices to unlock deep personalization.
  5. Create templates, notebooks, scripts, and repeatable workflows that improve how we analyze performance across customers.

Skills

Required

  • Strong ability to perform deep exploratory data analysis in Python (Polars / Pandas, Jupyter notebooks).
  • Ability to write and interpret SQL for customer warehouse analysis.
  • High-level understanding of ML modeling concepts (features, hyperparameters, reward functions, training windows).
  • Excellent communication skills; able to explain technical reasoning simply and confidently to marketers.
  • A customer-first attitude with high ownership and urgency when resolving issues.

Nice to have

  • Experience setting up and analyzing marketing experiments such as A/B, multivariate tests.
  • Prior experience in an applied ML, data science, analytics engineering, or forward-deployed role.
  • Experience building lightweight internal tools or scripting solutions.

What the JD emphasized

  • ensure that AI-driven marketing campaigns deliver measurable, compounding impact
  • make sure that these AI agents perform at their best
  • explain _why_ AI Decisioning is driving lift
  • debug performance issues, iterate on reward functions
  • explain technical reasoning simply and confidently to marketers

Other signals

  • AI agents for marketing workflows
  • AI Decisioning continuously learns customer preferences and executes 1:1 messaging
  • ensure that these AI agents perform at their best
  • debug performance issues, iterate on reward functions
  • explain how the decision engine handles cold start, message transfer learning, exploration vs. exploitation