SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Higgsfield AI is a rapidly scaling generative AI company with 25M+ users, 6M+ generations per day, and $500M+ annual revenue run rate, powering 390 Fortune 500 brands. This role owns the machine learning systems that personalize the product experience and monetization strategy for each user.
You will build and ship production ML models across three core areas:
**Recommendation & Ranking Systems**: Design ranking systems for content surfaces (effects, presets, templates, models, prompts) that optimize for user-specific value rather than average behavior. Solve cold-start problems for new users with minimal onboarding signals. Address real-world challenges including position bias, popularity feedback loops, and exploration-exploitation tradeoffs. Optimize for completed, kept, and shared generations rather than clicks.
**Uplift Modeling & Personalized Offers**: Own incrementality on offers (discounts, vouchers, trials, plan upgrades, win-back campaigns). Build models that identify users whose decisions are actually changed by offers, not just those likely to convert anyway. Design and maintain randomized holdouts for permanent measurement. Incorporate margin and abuse constraints into optimization objectives. Collaborate with Legal on personalized pricing compliance across markets.
**Churn, Retention & Lifecycle Propensity**: Build churn and downgrade prediction for subscribers and repeat-purchase propensity for credit buyers. Distinguish between models that predict churn and models that reduce it. Detect and eliminate feature leakage. Pair every propensity model with interventions and experiments.
**Generation Intelligence**: Extract features from user-generated content including use-case labels, intent classification, and multimodal understanding. Mine failed and abandoned generations as churn and unmet-demand signals. Build taxonomies that enable downstream ranking and offer models.
You own models end-to-end: problem framing, feature engineering, training, offline evaluation, serving, monitoring, and retraining. All work ships behind experiments with online validation. You track and maintain the business value of each deployed model.
Required: Proven experience shipping ML systems that changed live business metrics (not notebooks or benchmarks). Deep expertise in at least two of: recommender systems/learning-to-rank, uplift and causal ML, churn/propensity modeling, or real-time personalization. Strong causal reasoning (randomized holdouts, incrementality, Qini evaluation, selection effects). Proficiency in Python and SQL, comfort with gradient boosting and neural ranking. Product judgment to frame decisions and metrics before selecting models. Pragmatism to ship heuristics quickly and iterate. B2+ English proficiency.
