Staff Data Scientist, Forward Deployed

Forward Deployed Engineering · Predictive track · Staff

TEAM
Forward Deployed Engineering (FDE), Predictive track, in Data Science & AI Engineering

REPORTS TO
Head of Data Science & AI Engineering

LEVEL
Staff (10+ years) — leads the team's customer engagements

LOCATION
United States

THE ROLE IN ONE PARAGRAPH

Our largest customers want data scientists who work directly with their teams. As Staff Data Scientist, Forward Deployed, you'll lead that work: partnering with strategic customers to tailor our predictive models to how they go to market, and making sure the results show up in their pipeline.

You'll set the technical direction for five to ten customer engagements at a time while staying hands-on, and carry what you learn back into the product. This is a high-ownership role for someone who enjoys ambiguity, likes being close to customers, and wants their work to show up both in customer results and in the product.

What You'll Do

Build for the customer

• Lead the technical side of strategic customer engagements, from discovery to a solution running in production.
• Partner with customers' marketing and revenue operations teams to understand their go-to-market, data and goals, and find where predictive intelligence can have the most impact.
• Tailor and optimize models for each customer: ideal customer profile and industry definitions, keyword and intent strategy, how CRM and marketing automation activity feeds the models, and score thresholds.
• Shape each engagement with Professional Services and Customer Success: scope, plan, estimate and success measures.
• Lead customer reviews that recommend the next model, backed by backtests and conversion analysis customers can verify in their own data.
• Tell the story of the work in the customer's language, so results land with business leaders as well as practitioners.
• Guide customers through model and configuration changes so their teams know what to expect.

Productize for everyone

• Turn what works into reusable patterns, and bring the ones that recur across customers to Product as roadmap input backed by field evidence.
• Set the team's standards for discovery, validation, documentation and handoff, and mentor its data scientists.

What You'll Bring

• Required: 10+ years building and deploying machine learning models in production, ideally supervised models on tabular or behavioral data (propensity, scoring, ranking or similar).
• Experience working directly with customers on their data: consulting, client-site work, or a forward-deployed or solutions role at a company like Palantir or in fintech.
• Strong applied ML fundamentals, including tree-based models such as XGBoost and Random Forest: feature engineering, evaluation, calibration, threshold setting, and healthy skepticism about what a metric is telling you.
• Experience owning model changes in production end to end: backtesting, promotion decisions and rollback.
• Fluent in Python and SQL, and comfortable in messy, fragmented customer data rather than clean benchmarks.
• Able to run two or three engagements at once and switch between them without a fixed sprint plan.
• Exceptional communication: you can tell the story of the work to an executive and to an engineer.
• A product mindset: you'd rather build a capability once for every customer than solve it ten times by hand.
• Nice to have: B2B go-to-market data (CRM, marketing automation, intent, ABM or predictive scoring); explaining model output to non-technical users; leading a small team or a portfolio of client engagements; building or evaluating LLM-based tools and agents; data platforms such as SingleStore, Hive or Spark on AWS.

What Success Looks Like

• In 3 months: you've led engagements end to end and delivered a measurable improvement in at least one enterprise customer's predictive results.
• In 6 months: patterns from your engagements are on the predictive roadmap, and at least one tailored solution is on its way to becoming a standard product capability.
• In 12 months: engagements move faster because of the tooling and playbook you built, and more of what customers need comes in the product out of the box.

Why this role is different: most data science roles build models at a distance from the customer or advise without shipping. This one leads the work end to end, from the customer conversation to the model in production, and turns it into product.