SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
CarOnSale is an AI-powered B2B platform for used car trading across Europe, connecting 40,000+ buyers from 20+ countries with exclusive inventory. The company operates as an integrated OS for the automotive trading industry, combining software, pricing intelligence, logistics, and financing.
You will own the full lifecycle of machine learning models from handoff through production deployment and monitoring. The company is scaling from four production models today to 15–20 by mid-2027, all running on a shared, centralized ML platform rather than separate pipelines. Your core responsibilities include:
• Model packaging, deployment, and serving: you own the inference path, fitted pipeline artifacts, monitoring hooks, and feature-store alignment
• Production reliability: drift detection, performance monitoring, alerting, and incident response
• Quality gates: you review model design and evaluation methodology before production, catching data leakage, backward-window errors, and weak evaluation early
• Platform evolution: extend the shared ML platform to support new models without project-specific logic leaking into shared code
• Engineering standards: set and maintain the standards the platform runs on as it scales
The platform is built on Snowflake (data warehouse), SageMaker (managed feature store), and AWS infrastructure managed via Terraform, with governed CI/CD promotion lanes.
You bring 2+ years of production ML engineering with real post-training ownership, strong typed and tested Python, enough ML depth to challenge problem framing and feature engineering, hands-on experience with managed ML platforms (SageMaker, Vertex AI, Databricks, or Azure ML), and active use of AI tools (Claude, ChatGPT, Copilot) in daily work. C1 English required; German not required. Nice-to-haves include Snowflake and dbt experience, mentoring experience, and comfort in ambiguity.
Hybrid role: 3 days office in Berlin Schöneberg, 2 days remote, plus 25 "Work from Anywhere" days annually. Benefits include 28 days leave, 20% pension contribution, Deutschlandticket, fitness subsidy, virtual stock options, and structured onboarding.
