data scientist for subscription monetization

Higgsfield AI develops AI-powered video creation and next-generation creative tools. The company reports a $1B annual revenue run rate, more than 25 million users, over 6 million generations per day, and work with 390 Fortune 500 brands.

• Build separate churn and downgrade prediction models for subscribers and repeat-purchase propensity models for one-time credit-pack buyers
• Pair each score with an intervention and a holdout, focusing on retained users rather than ranked lists
• Own incrementality measurement for discounts, credit vouchers, trials, upgrade prompts, win-back campaigns, and retention saves
• Design randomized holdouts for uplift training and measurement, and keep them running
• Optimize offer targeting within margin and abuse constraints
• Work with Legal on personalization, consumer-protection, and consent rules
• Own cohort-based LTV measurement using contribution margin, segmented by plan, segment, and acquisition discount
• Model refunds, chargebacks, credit breakage and expiry, plan migration and downgrades, and compute usage
• Use LTV to inform user acquisition costs, discount depth, and plan decisions; keep LTV current and versioned
• Design and analyze experiments for pricing, packaging, offers, and lifecycle with the product analytics team
• Use causal methods when clean experiments are not possible, and state their assumptions
• Distinguish correlation from causal effects and rigorously validate findings
• Own model delivery end to end, from decision framing and features through training, evaluation, deployment, monitoring, and retraining
• Ship batch scores and policy rules into lifecycle tooling, the paywall, and pricing, with readouts of their financial value
• Monitor model drift and keep each model’s financial value current

• Experience shipping models that changed business actions toward real users, such as targeted campaigns, pricing rules, or retention programs
• Depth in at least two areas: churn and propensity modelling, uplift and causal ML, LTV and subscription economics, or experimentation at scale
• Strong causal inference skills, including randomized holdouts, incrementality, Qini and uplift curves, and selection effects
• Strong SQL and Python skills, including cohort, funnel, and retention analysis on raw event data, gradient boosting, feature pipelines, retraining cadence, and scheduled scoring jobs
• Understand subscription and one-time purchase mechanics, including recurring and one-off revenue, refunds, plan changes, deferred value of unspent credits, and revenue versus margin
• Apply commercial judgment to define decisions and metrics before choosing a model, and recognize when a rule is more appropriate
• Work pragmatically, shipping useful solutions and replacing them when needed
• Clear written and spoken English, B2+
• Not a fit for candidates who want to train or fine-tune generative video models, own a recommender or serving stack, target users based only on conversion propensity, report misleading lift, optimize AUC without owning the resulting action, require mature data infrastructure before starting, or want predictable 9–5 workdays
• Будет плюсом: experience in growth, retention, or monetization data science in subscription, gaming, fintech, or e-commerce; CRM and lifecycle campaign targeting and incrementality; causal inference and uplift policy deployment; product analytics experience combined with modelling

• Equity participation through the company’s stock option program
• Relocation support to Almaty for candidates moving from another city or country
• Company-provided equipment, meals, transportation, or other office benefits
• Paid on-site trial for 1 month
• Practical home task with a 7-day timeframe