Data Science

Playtika is looking for a Data Scientist  – Production ML to join the Data & AI department 

In this position you will join a multidisciplinary team focused on personalization using Reinforcement Learning methods. Over the last few years, we have built a real-time recommendation engine based on Bayesian Multi-Armed Bandits - including our open-source PyBandits library - serving millions of players across multiple studios and use-cases. 

This is an applied research role, with impact measured in production. Your primary contribution will be advancing the mathematical and statistical foundations of our existing solutions: identifying limitations, deriving principled improvements, and owning those improvements all the way through to production. You will not be handed research problems from above - you will find them yourself, inside systems that are already running at scale.  

Scaling here is fundamentally a mathematical problem: identifying better statistical methods, proving they hold under real-world constraints, and knowing exactly how far they deviate from the ideal. That's what sets this role apart. 

 

Responsibilities:

• Identify mathematical and statistical limitations in existing ML solutions and deliver provably correct, more efficient alternatives 
• Validate the correctness of scaled solutions - diagnosing where distributed execution or approximation invalidates prototype-level results, and redesigning accordingly
• Own the full lifecycle of your improvements: from research and prototyping through production deployment and impact measurement  
• Drive the evolution and scalability of our personalization and recommendation systems to meet future business and technical requirements  
• Contribute to end-to-end data-driven research including problem definition, data collection, model development, evaluation, and deployment
• Collaborate with cross-functional teams (engineers, BI, Product) to deploy solutions quickly and effectively
• Mentor Data Scientists and ML Engineers, and serve as a technical reference on statistical and mathematical rigor across the organization  

 

 

Requirements:  

• M.Sc or PhD in Computer Science, Mathematics, Statistics, Engineering or a related field
• 5+ years of relevant working experience, including significant production ML ownership
• Demonstrated experience taking statistically sophisticated models from prototype to production and operating them successfully under real-world business constraints.  
• Experience with neural network architectures, including the ability to diagnose, interpret and debug their behavior in probabilistic or decision-making contexts (e.g. Bayesian Neural Networks, Neural Linear models)
• Strong theoretical foundation in at least one of: Multi-Armed Bandits, Bayesian methods, online learning, recommendation systems, or related decision-making frameworks
• Strong understanding of experimentation, causal inference and statistical evaluation of online decision systems.
• Strong hands-on Python engineering skills, with the ability to design, implement, review, and maintain production-quality code, leveraging modern AI coding assistants (e.g., Claude, Cursor) to improve productivity and code quality
• Proficiency in SQL
• Clear written and verbal communication - able to explain statistical reasoning to non-technical stakeholders
• Ability to operate autonomously and influence decisions across teams without direct authority. 

 

Nice to have 

• Experience with probabilistic programming frameworks (e.g. PyMC or Numpyro)
• Familiarity with MLOps tooling (experiment tracking, model serving, monitoring)
• Familiarity with approximate inference methods (variational inference, MCMC)
• Experience with PySpark
• Experience in the Mobile Gaming industry