MLOps Engineer

What You ll Do

• Build and operate end-to-end ML/LLM pipelines: data ingestion, feature processing, training, evaluation, packaging, registry and deployment.

• Automate workflows: design fault-tolerant training/inference pipelines with Kubeflow; implement CI/CD for ML with GitHub Actions and reusable templates.

• Serve models: containerize ML model s (Docker), expose APIs ( FastAPI ), deployments in Google Cloud Ve rtex AI and Kubernetes.

• Ensure observability and monitoring: implement metrics, logs and traces; set up model/data quality checks, drift detection and alerting

• Optimize cloud cost and performance: finetune the usage of compute resources and apply cloud best practices.

• Collaborate: work with ML Engineers and Data ( Ops ) Engineers to deliver quality products; review code and doc umentation ; apply best coding practices for maintainable and reusable code; support junior colleagues to grow the MLOps capabilities.

• Contribute to our culture: bring an experimentation mindset, propose improvements, and help us stay current with modern MLOps tooling and practices.

• Coach and mentor more junior team members, ensuring that MLOps skills and capabilities are developed at scale

M ust-Have Qualifications

• Education: BSc or MSc in Math, Physical Sciences, Computer Science, Software Engineering, AI/ ML or related field .

• Software: Experienced Python knowledge , testing practices, Git/GitHub , GitHub Actions , Docker; experience building APIs with FastAPI or similar.

• Cloud: hands-on experience with at least one major provider (GCP/AWS/Azure) and core services ( compute , storage, networking , AI platforms ) .

• MLOps : experience with pipeline orchestration (Airflow/ K ubeflow), experiment tracking/model registry .

• Monitoring: practical experience setting up dashboards and alerts (e.g., Prometheus/Grafana/ OpenTelemetry ) and ML-specific monitoring for data drift and performance.

• Frameworks: PyTorch or TensorFlow in production contexts.

• Communication: clear, proactive communicator in English; able to collaborate with diverse stakeholders.

• Work mode: open to hybrid work; team typically spends ~2 days/week in the office.

Nice to Have

• LLMOps : prompt and experiment tracking (e.g., Langfuse ), evaluation frameworks, guardrails, vector databases ( e.g. Pinecone).

• Infrastructure as Code and packaging: Terraform / Pulumi

• Experience or understanding of deploying in Kubernetes

• Inference optimization: model quantization, Triton/ vLLM / TensorRT , GPU operators and scheduling.