Machine Learning Engineer

Oak Tree Software; Deutschland; Vor Ort; EUR 90.000 - 130.000; Vollzeit

Zusammenfassung

Oak Tree Software is seeking an experienced ML/ML Ops Engineer to build and operate production-grade ML infrastructure in a Germany-based role. You will implement end-to-end pipelines with Kubeflow, manage GPU resources in Kubernetes, and fine-tune LLMs while tracking experiments in MLflow.

You’ll work with SQL Server and DuckDB for fast in-cluster data processing, coding in Python, and adhering to zero-trust security policies in enterprise cloud environments.

Qualifikationen

• Hands-on experience with Kubeflow Pipelines (KFP v2).
• Experience training models on GPUs.
• Experience fine-tuning LLMs/transformers.
• Experience with MLflow (model tracking).
• Experience with boosting models (XGBoost, CatBoost).
• Experience with SQL and large-scale data processing.
• CI/CD experience (GitLab CI preferred), clean code and testing practices.

Aufgaben

• Build and orchestrate ML pipelines using Kubeflow Pipelines (KFP v2).
• Train models on GPUs, including GPU resource management within Kubernetes.
• Fine-tune transformers/LLMs, track experiments and models via MLflow.
• Build classic ML models (XGBoost, CatBoost).
• Work with data using SQL Server and DuckDB as a lightweight OLAP solution for efficient in-cluster processing of large datasets.
• Develop in Python (pipelines, integrations, tooling based on uv).
• Work within a zero-trust / secure-by-default environment (network policies, restrictive container rights).

Kenntnisse

Kubeflow Pipelines

GPUs training

LLM fine-tuning

MLflow

XGBoost

CatBoost

SQL/large-scale data

GitLab CI

Clean code/testing

Tools

Python

Kubernetes

DuckDB

uv

Jobbeschreibung

Build and orchestrate ML pipelines using Kubeflow Pipelines (KFP v2)

Train models on GPUs, including GPU resource management within Kubernetes

Fine-tune transformers/LLMs, track experiments and models via MLflow

Build classic ML models (XGBoost, CatBoost)

Work with data using SQL Server and DuckDB as a lightweight OLAP solution for efficient in-cluster processing of large datasets

Develop in Python (pipelines, integrations, tooling based on uv)

Work within a zero-trust / secure-by-default environment (network policies, restrictive container rights)

Candidate's Portrait

Location: Germany

German language at B2 level is a must have

An experienced ML/ML Ops Engineer with a strong engineering background, combining:

Hands-on, production-grade ML infrastructure experience (not just notebook-level ML)

Deep proficiency in the Python ecosystem and modern engineering practices

Experience working in regulated/enterprise cloud-native environments

Willingness to quickly ramp up on the client's domain (healthcare/billing data), even without prior background in it

Must-haves

• Hands-on experience with Kubeflow Pipelines (KFP v2)
• Experience training models on GPUs
• Experience fine-tuning LLMs/transformers
• Experience with MLflow (model tracking)
• Experience with boosting models (XGBoost, CatBoost)
• Experience with SQL and large-scale data processing
• CI/CD experience (GitLab CI preferred), clean code and testing practices

Nice-to-have

• Pre-training experience for LLMs (beyond fine-tuning)
• Experience in zero-trust environments (network policies, restrictive container rights)
• Knowledge of DuckDB
• Experience with modern Python tooling (uv)