8 hours ago
Berlin, GermanySenior
Responsibilities
- Migrate existing model training code from experimental or proof-of-concept stages into robust, automated pipelines.
- Design, build, and maintain continuous training workflows covering data preparation, model training, and batch scoring.
- Package and deploy traditional ML models, including ONNX models, to serving infrastructure such as Vespa and Triton.
- Automate model deployments, including deployments triggered by successful Vertex AI training pipelines through GitHub Actions.
- Build automated model quality and evaluation gates for ML pipelines.
- Own and continuously improve standards, quality, and governance across the complete ML model lifecycle.
Requirements
- Strong software engineering background with an ML-literate, automation-focused approach rather than a Data Scientist profile.
- Deep expertise in GCP/Vertex AI or equivalent platforms such as SageMaker.
- Experience with Kubernetes and Docker.
- Proven experience building robust CI/CD and automation pipelines using tools such as GitHub Actions or Jenkins.
- Understanding of ML training, evaluation, and inference workflows sufficient to debug pipelines and build meaningful quality gates.
- Strong service mindset, communication skills, pragmatism, problem-solving ability, and capacity to bridge Data Science and Infrastructure teams.
- Familiarity with Vespa, NVIDIA Triton, and ONNX is a bonus.
- Willingness to learn and master new technologies and concepts.
Benefits
- Share options programme.
- 30 days of paid annual leave.
- Newest MacBook models.
- Confidential Employee Assistance Program for employees and their families.
- Home-office IT workstation equipment and up to €540 for home workplace furniture.
- Lunch benefit, team-building events, monthly Vinted shopping budget, discounted gym membership, pension matching at 150% of the chosen contribution, supplemental private health insurance, life and disability insurance, and a subsidised Deutschlandticket.
- Hybrid work with two recommended office days per week.
- Workation policy allowing remote work for up to 90 days per year in the EU, including up to 21 days globally; non-EU citizens may work remotely up to 21 days worldwide.
- Annual learning budget starting at €3,000 and up to 10 working days over two years for professional development.
- Dog-friendly office.
Tech Stack
Categories
About Vinted
Vinted runs a consumer-to-consumer marketplace and app for buying and selling second-hand fashion and related items across Europe. It monetizes through transaction-related fees, buyer protection, and shipping services, and also operates Vinted Go for logistics and Vinted Pay for payments. Founded in 2008 and headquartered in Vilnius, Lithuania, the platform serves members in 20+ European markets.