ML Ops Enigneer / 1

Job Description

We are seeking an advanced ML Ops Engineer to design and implement the infrastructure required to host, orchestrate, and manage up to 1,500 ML scoring processes within a new Databricks environment. The focus of the role is on operationalizing the ML scoring pipelines by setting up a scalable, secure, and well‑monitored platform for data science teams to deploy their models.

Qualifications

Environment Configuration

• Set up Databricks clusters, jobs, and workflows for large-scale ML scoring use cases.
• Infrastructure as Code is used for reproducibility and governance (e.g., Terraform).
• Implement scalable infrastructure capable of running thousands of ML scoring tasks.
• Configure job scheduling, parallel execution strategies, and resource optimization.
• Monitoring and alerting are integrated into the platform using cloud-native tools.
• Security, compliance, and cost-efficiency are key pillars of the operational setup.

ML Ops Pipeline Integration

• Develop deployment processes for ML models using Databricks MLflow or equivalent.
• Implement version control and tracking for models, scoring code, and configuration files.

Execution Management

• Build frameworks to orchestrate scoring of >1,500 ML models or scoring jobs.
• Ensure resilience, fault tolerance, and restart capabilities for failed jobs.
• Monitoring & Observability Integrate logging, alerting, and dashboards to monitor scoring throughput, latency, and failures.
• Establish model performance monitoring hooks for post‑scoring analytics.

Automation

• Work alongside Dev Ops Engineers to ensure common infrastructure and processes (e.g., shared storage, Delta Lake tables) serve both ML and BI use cases.
• Automate provisioning of resources and deployments from CI/CD pipelines.
• Utilize Infrastructure as Code (IaC) where feasible for reproducibility.

Collaboration

• Work closely with data scientists, solution architects, and platform engineers to ensure smooth handover from model development to operational scoring.
• Define operational SLAs for scoring workloads.

Additional Information

Work 3 times a week from an office in Warsaw, Lublin or Poznań.

We hereby inform you that Inetum Polska sp. z o.o. has implemented an internal reporting (whistleblowing) procedure. The content of the procedure and the possibility to submit an internal report are available at:

https://inetum.whispli.com/speakup?locale=pl