Principal Applied Scientist

About the Company

Playvalve is a vibrant fast-growing gaming studio with an eye on the future. Our mission is to build fun and relaxing games meant to last, and we do that by bringing together a team of experts who love what they do. Our talented designers, artists, engineers, marketeers and analysts have created several games delivering joy and entertainment to over 20M users across more than 12 languages and 90 countries.

About the Role

As a Principal Applied Scientist, you'll bridge rigorous science with production impact. This is not a pure engineering role, we need someone who thinks like a scientist first: formulating hypotheses, designing experiments with statistical rigor, and building models grounded in solid mathematical foundations. You'll then work with Engineering to productionize these solutions at scale.

You'll own end-to-end problems in user behavior modeling, ad monetization optimization, and real-time decisioning systems. This means deep work on causal inference, probabilistic modeling, experimentation design, and optimization, not just applying off-the-shelf ML libraries. You'll need to understand the 'why' behind the models, not just the 'how.'

We're looking for someone who holds an advanced degree in a quantitative field, and has battle-tested their science in production environments. You should be equally comfortable working through the math behind a model as you are debugging production pipelines.

Key Responsibilities

Scientific Leadership & Problem Framing (50%)

• Frame ambiguous business problems as well-defined scientific questions with testable hypotheses.
• Design and analyze controlled experiments (A/B tests, multi-armed bandits) with proper statistical rigor.
• Build probabilistic models for user behavior, lifetime value prediction, and ad response.
• Apply causal inference techniques to untangle correlation from causation in observational data.
• Stay current with research literature and bring cutting-edge methods to production (when appropriate).

Production ML Systems (20%)

• Partner with Engineering to build real-time decisioning APIs for game configuration and ad serving.
• Implement proper monitoring, evaluation frameworks, and fallback mechanisms.
• Ship models as stable services with clear performance guarantees.

Experimentation & Measurement (20%)

• Define success metrics and design experiments that provide unambiguous answers.
• Build experimentation infrastructure and champion rigorous hypothesis testing across the org.
• Quantify uncertainty and communicate statistical trade-offs to non-technical stakeholders.
• Translate experimental results into business decisions with estimated impact.

Strategic Collaboration & Mentorship (10%)

• Work with Product, Marketing, and Ad Monetization to prioritize high-impact projects.
• Mentor other data scientists on scientific rigor.
• Establish standards for model evaluation, code quality, and documentation.
• Communicate complex technical concepts clearly to diverse audiences.

Requirements

• PhD or MSc in Computer Science, Statistics, Applied Mathematics, Physics, Operations Research, or closely related quantitative field.
• 7+ years in data science with 2+ years owning production ML systems.
• Expert-level Python and SQL; comfortable writing production-quality code.
• Deep understanding of probability theory, statistical inference, and mathematical optimization.
• Strong grasp of experimentation design and statistical hypothesis testing (power analysis, multiple testing corrections, variance reduction techniques).
• Demonstrated ability to read, critique, and apply research papers to real-world problems.

Nice to have

• Experience with modern ML frameworks (TensorFlow, PyTorch, JAX, or similar) for deep learning and/or probabilistic programming (e.g., PyMC, Stan, TensorFlow Probability).
• Built and deployed end-to-end ML systems that handle millions of requests/predictions.
• Proficiency with cloud platforms (GCP, AWS, Azure) for data processing and model serving.
• Published research in top-tier venues (NeurIPS, ICML, KDD, JMLR, etc.) or industry conferences.
• Relevant Domain Experience (1-2 of the following strongly preferred):
  • Gaming/mobile apps: LTV prediction, retention modeling, monetization optimization
  • Ad tech: Bidding strategies, auction mechanisms, ad serving optimization
  • Marketplace/platforms: Two-sided market dynamics, pricing, recommendations.
  • Subscription/SaaS: Churn prediction, upsell optimization, cohort analysis

What we offer

• The Role: Own data science solutions from scratch with full autonomy. Join a meetingless company, no status updates, no sync calls, just deep work and results.
• Location & Flexibility: Based in Barcelona (office with terrace, BBQ, beer tap). Flexibility policy: the data team comes in regularly because they want to, not because they have to.
• Compensation & Perks: Competitive compensation based on experience, phantom stocks, relocation package, health insurance, ticket restaurant, unlimited vacations, regular company off-sites.
• The Challenge: 15 games in market + ambitious pipeline. Industry inflection point with AI and automation. Profit-backed studio with no VC pressure and resources to do it right.