Senior MLOps Engineer

Quick Facts

• Role: Senior MLOps Engineer

Description

Design, build, and maintain scalable MLOps and LLMOps infrastructure on GCP. Develop and manage ML pipelines, containerized environments, CI/CD workflows, and LLM/agentic orchestration (including evaluation, monitoring, and observability) to deliver production-ready AI solutions.

Responsibilities

• Design, build, and maintain scalable MLOps and LLMOps infrastructure on GCP

• Develop and manage ML pipelines using MLflow, ZenML, Kubeflow, and Vertex AI Pipelines

• Manage containerized environments with Docker and Kubernetes

• Automate infrastructure provisioning with Terraform

• Build and maintain CI/CD pipelines for ML and GenAI apps with GitHub Actions

• Orchestrate LLM and agentic workflows using LangGraph

• Implement evaluation frameworks for LLMs, RAG systems, and AI agents

• Monitor ML/GenAI applications for performance, drift, tracing, and observability

• Develop scalable data processing pipelines with Dataflow and Apache Beam

• Collaborate with Data Scientists and Engineering teams to deliver production-ready AI solutions

• Provide technical guidance and promote MLOps/LLMOps best practices across the team

Requirements

• Strong professional experience in MLOps/LLMOps

• Hands-on experience with GCP (especially Vertex AI, Dataflow, BigQuery)

• Strong knowledge of Terraform, Docker, and Kubernetes

• Experience with Kubeflow, Argo Workflows, MLflow, ZenML, and/or Vertex AI Pipelines

• Strong Python and Bash skills

• Experience building CI/CD pipelines for ML and/or GenAI applications

• Practical experience with Generative AI and LLM applications

• Experience with LangGraph or similar AI orchestration frameworks

• Knowledge of LLM evaluation, RAG evaluation, model monitoring, and drift detection

• Familiarity with Kafka, Spark, SQL/NoSQL databases, and Data Lakes

• Communication, problem-solving, and mentoring skills

• Spanish: B2 minimum (C1 preferred)

• English: B2 minimum

Benefits

Not specified in the provided posting.