Backend Engineer - Recommendation (all genders)

Description

The company runs an online fashion shop and is one of Europe's fastest growing e-commerce companies, with offices in Hamburg and Berlin.

The team works on personal shopping, which means the company is trying to rethink online shopping, with flat hierarchies, direct communication and quick decisions.

The role sits on the recommendation engine behind the homepage and the product and outfit pages, and is a backend role at the point where ML and live systems meet.

Responsibilities

• Build and run the recommendation engine behind the homepage and product and outfit pages, written in Node.js and TypeScript.
• Serve offline-trained models (matrix factorization, brand similarity, image embeddings) at Black Friday peak load from Valkey, DynamoDB and OpenSearch.
• Design the personalization layer: feed assembly, fallback chains, deduplication, diversity and A/B experiments.
• Develop GenAI features such as the streaming Gemini fashion advisor, focusing on latency, session state and token cost.
• Own reverse-ETL pipelines that move BigQuery model outputs into fast hot stores.
• Track recommendation coverage and business metrics through SQS and BI systems, and help shape the architecture with the Tech Lead.

Requirements

• You own the recommendation engine end to end and want to see your changes affect customer engagement directly.
• At least 3 years building high-throughput backend services. Node.js and TypeScript preferred; engineers coming from Java, Go or Kotlin who can switch quickly are welcome.
• Solid experience with Redis or Valkey data modeling, multi-tier caching, TTLs and invalidation, and binary formats such as Protobuf or Avro.
• Comfortable running APIs at 10k+ requests per second under strict latency targets.
• Interest in bridging backend and ML by serving model scores, embeddings or LLM features in production.
• Clear English communication and a data-driven, experiment-minded approach.
• Knowledge of recommender systems: collaborative filtering, matrix factorization, cold-start handling and diversity algorithms.
• Experience with streaming GenAI or LLM products, including chat UX, prompt caching and guardrails.
• Experience with AWS serverless (DynamoDB data modeling, SQS and SNS) and OpenSearch k-NN indices.
• Experience profiling performance (clinic.js, flamegraphs) and load testing with k6.

Conditions

Applications are made online through the company career page.

The company is committed to an inclusive workplace where everyone can feel welcome, and it sees diverse backgrounds as a strength.

Benefits

• Perks are listed on the company benefits page.