Machine Learning Engineer (Berlin)

About the Team

The Performance team is the brain behind our bidding infrastructure. We handle the core algorithms responsible for maximizing ad performance and optimizing return on ad spend (ROAS) within the strict parameters set by our advertisers. We run a hybrid setup; training in the cloud and inference on our own hardware. That mix gives us the freedom to push the boundaries of high-throughput ML, while keeping us honest about the cost and scale trade-offs of doing so. We have a pragmatic view about what 'ML' means here; feature pipelines, online serving, and monitoring are part of the job, not someone else's problem.

In this role you will...

• Develop & Scale ML Models: Build, train, and deploy high-performance machine learning prediction models optimized for real-time adtech environments.
• Own the Full ML Path: Take models from raw data and feature definitions through training, evaluation, and online serving. Feature engineering and online/offline consistency matter as much as modelling skill.
• Optimize the Bidding Engine: Design and refine complex bidding algorithms, orchestrating how multiple models interact together in milliseconds to make optimal final bidding decisions.
• Run & Analyze Experiments: Design, execute, and evaluate both offline simulations and online A/B tests to measure and analyze algorithmic impact on core business metrics.
• Collaborate Cross-Functionally: Step outside the traditional ML silo to collaborate with engineering and product partners, adapting quickly to diverse technical challenges.
• Augment Your Workflow: Actively integrate new AI tools and coding agents into your daily process to maximize throughput, experiment velocity, and code quality.

You will be a great fit if...

• Technical Stack: You have strong proficiency in Python and TensorFlow for building ML models, and comfortable reading Go to navigate our online serving services, this would be easy to pick up if you already know Python.
• Growth Mindset: Beyond core ML, our work touches a range of disciplines like reinforcement learning, control engineering, mathematical optimisation, and causal inference, to name a few. If you already have depth in any of these, great; if not, there's plenty of interesting territory to explore alongside the core ML work.
• Healthy Skepticism & Pragmatism: You don't just build complex models because they are trendy; you use data-driven analysis to find the most pragmatic solution to the problem at hand.
• Ownership Mindset: You are adaptable and thrive when given the autonomy to take a project from early-stage experimentation all the way through to production.
• AI-Forward: You possess a genuine curiosity about the evolution of AI and are enthusiastic about leveraging AI coding tools to optimize your engineering workflows.