(Senior) Data Scientist

• Your recommendation engine suggests a product category - opening up €50K in new monthly revenue
• Your clustering algorithm reveals that 20 customers have untapped potential worth €5M

Welcome to wholesale AI - where sparse data meets high stakes, where relationships matter more than transactions, and where the right insight at the right time can transform a business., * Build and deploy ML models processing millions of transactions across 20+ enterprise customers

• Design recommendation engines handling catalogs of 500K+ products
• Develop predictive models for customer behavior with sparse, irregular interaction patterns
• Create churn prediction and cross-sell systems integrated with ERP systems
• Own complete lifecycle from problem definition to production deployment
• Build evaluation frameworks connecting model performance to business KPIs
• Explanations matter as much as predictions

What you'll be working on

Production AI Systems at Scale

• Build and deploy ML models that process millions of transactions across 20+ (and growing) enterprise customers simultaneously
• Design systems that adapt to wildly different data distributions (a construction wholesaler vs. a medical equipment distributor)
• Create recommendation engines that handle catalogs of 500K+ products where most customers buy <5% of items
• Develop predictive models for customer behavior with sparse, irregular interaction patterns
• Design churn prediction and cross-sell systems that integrate seamlessly with ERP (Entereprise Resource Planning) systems
• Build models that generate automated proposals and personalized communication templates using LLMs.
• Combine classical ML (churn prediction, clustering) with LLMs for insight generation and sales enablement

End-to-End Ownership

• Own the complete lifecycle from problem definition to production deployment
• Build evaluation frameworks that connect model performance to business KPIs
• Design A/B testing infrastructure for continuous improvement
• Create feedback loops that learn from real-world outcomes

Complex Technical Challenges

• Handle extreme class imbalance (95%+ negative cases) while maintaining business value
• Build models that work with limited historical data (cold-start problems)
• Design architectures that scale from SMBs with 100 customers to enterprises with 100K customers
• Solve multi-objective optimization problems (maximize revenue while ensuring diversity)

Do you have experience in Unity?, Do you have a Master's degree?, * Strong background in machine learning fundamentals - you understand why algorithms work, not just how to use them

• Experience building production data powered products systems that handle real-world messiness
• Familiarity in distributed computing (Spark/PySpark) for processing large-scale data
• Solid software engineering practices - your code is tested, documented, and maintainable

Problem-Solving Mindset

• You approach problems from first principles rather than reaching for standard solutions
• Comfortable with ambiguity - you can define success metrics when requirements are vague
• You balance technical elegance with business pragmatism
• Experience translating business problems into data science solutions

Proven Track Record In:

• Building ML systems that handle irregular patterns (time series with gaps, seasonal businesses, etc.)
• Working with hierarchical data structures (product taxonomies, customer segments)
• Creating models that provide actionable insights, not just predictions
• Deploying ML in multi-tenant architectures where one model serves many clients

Bonus Points For:

• Experience in B2B analytics, e-commerce, or supply chain optimization
• Knowledge of recommendation systems, customer analytics, or revenue optimization
• Familiarity with modern data platforms (Databricks, Snowflake, etc.)
• Experience with MLOps practices and model lifecycle management
• Understanding of European business practices and regulations
• Experience with LLM APIs, prompt engineering, or building LLM-augmented products
• Proficiency in German

Our Tech Stack

You'll be working with modern tools, but we care more about your ability to learn than specific tool experience:

• Data Processing: PySpark, SQL, Python
• ML Platform: Databricks (Unity Catalog, Workflows, Model Serving)
• ML Libraries: scikit-learn, XGBoost/LightGBM, implicit, scipy, OpenAI + experience with your preferred frameworks
• Infrastructure: AWS, Terraform
• Orchestration: Databricks Workflows, Github Actions
• Experimentation: MLflow

Plato is the AI-powered Sales Intelligence platform built specifically for B2B wholesale. We transform reactive sales teams into proactive revenue engines by surfacing hidden opportunities in their data. Our platform identifies which customers are at risk, who's ready to buy more, and what products to recommend next - turning thousands of signals into clear next actions. The Wholesale Challenge Why wholesale is different from everything else you've worked on: Imagine building ML for a world where: * Your "best customer" might buy twice a year - but spends €500K each time * A customer churning doesn't mean they stopped needing your products - they just found another supplier * Product catalogs contain 100K+ SKUs but even your best customers only know 500 of them * A "trend" might be 3 data points over 18 months * Success means helping a sales rep know which of their 200 customers needs attention today