Machine Learning Product Engineer

Menu Personalization determines what millions of customers see when they open HelloFresh each week. As a Senior Machine Learning Product Engineer in this team you will shape how millions of customers discover and curate their weekly meals across our global digital platforms. You will handle the recommender systems that match customers to recipes across global markets, bringing together Data Scientists, Backend Engineers, Data Engineers, ML Engineers, and Product to take ideas from experiment to production. The work directly shapes customer experience and business growth: when personalization improves, customers find recipes they love faster, and HelloFresh becomes a stronger weekly habit.

This position is for a Senior Machine Learning (ML) Product Engineer for the Menu Personalization team to help build and operate the recommender stack running in production. In this role you will design, build, and operate ML systems across feature pipelines, training workflows, model serving, experimentation tooling, and the underlying infrastructure, holding end-to-end accountability for significant parts of the stack.

The role involves bringing a distinct point of view on how to improve personalization, backed by data and user evidence. You will partner with Data Scientists to transition models from notebooks to production, with Data Engineers on features and pipelines, with Backend Engineers on online inference paths, and with Product on future roadmaps. This role does not include people management responsibilities.

At HelloTech, flexibility and cross-functional collaboration are core to how we work. While this role is aligned to a specific Alliance, strong candidates may also be considered for opportunities across different teams or projects.

At HelloFresh we are moving away from a model where software developers just execute tickets toward one where product engineers are trusted to own customer problems. A Product Engineer takes a problem, forms a point of view, validates it with customers and data, and ships it using AI as a force multiplier.

• Build and operate data products and ML systems behind menu personalization, working hands-on across feature pipelines, training workflows, model serving, experimentation tooling, and infrastructure.
• Transition research and experiments into reliable production systems, partnering with Data Scientists on services that meet real latency, scalability, and observability requirements.
• Maintain accountability for significant components of the recommender stack, from design through deployment and ongoing operation.
• Operate deliverables, instrumenting and improving systems in production to ensure continuous refinement based on real-world performance.
• Contribute to the personalization roadmap with Product and Engineering, backing technical directions with data and user evidence.
• Raise the technical bar on the team through thorough code and design reviews, technical guidance, and setting an example for production ML craft.
• Work beyond your specialization when the problem demands it. Your specialization is your anchor, not your boundary.
• Operate what you build. You instrument, monitor, and improve your systems in production. Shipping is the beginning of the learning cycle, not the end.

• Hands-on experience working with AI tooling (e.g., Claude Code, Cursor, Copilot) beyond casual experimentation. You use AI agents every day. You have a practical sense of how the context you provide to AI tools shapes output quality, and how to set boundaries on AI-generated work.
• 5 years (ideally) of experience building and operating production ML systems.
• Fluency across the data and ML stack (Python, Apache Spark) and working knowledge of the backend and platform stack (Kafka, Kubernetes), with hands-on experience across pipelines, model serving, and observability at scale.
• Deep data/ML engineering expertise, experience operating models or data products in production, and the statistical literacy to design sound experiments and interpret their results.
• Production experience with recommender systems or large-scale personalization is a strong plus.
• Operational judgment to diagnose system misbehavior under real load, identify root causes, and deploy robust fixes.
• You take full ownership. You have a bias to ship. You finish the last twenty percent.
• You have product sense and are opinionated about what should be built and why, and you can back that opinion with data and u