MLOps Engineer

Role Overview

Build and maintain CI/CD and deployment infrastructure for machine learning models moving from research to production. Work at the intersection of ML engineering and DevOps to ensure reliable, low-latency, high-availability model delivery.

Responsibilities

• Design and implement automated CI/CD pipelines for ML model packaging, testing, and deployment (e.g., GitHub Actions, GitLab CI, Argo Workflows).
• Build and operate model orchestration and serving platforms using Kubernetes and ML-serving tooling (e.g., EKS/GKE, KServe, Triton, MLflow).
• Create real-time and batch feature store pipelines (e.g., Feast or Tecton) to connect offline training with online serving.
• Establish observability and monitoring for production models, including model health, latency, and data/concept drift (Prometheus, Grafana, Evidently, Arize).
• Optimize inference performance using techniques and accelerators such as quantization, pruning, and GPU/TPU/TensorRT.
• Integrate model deployments with data and streaming platforms, including data lakes/warehouses (Snowflake, Databricks, BigQuery) and streaming (Apache Kafka).
• Collaborate with data scientists, data engineers, and backend engineers on artifact management and infrastructure standards.

Requirements

• 3–6 years of professional experience as an MLOps/DevOps/Infrastructure Engineer supporting data science teams.
• Strong proficiency in Python and shell scripting.
• Deep experience with Docker and Kubernetes.
• Hands-on cloud experience with AWS or GCP.
• Infrastructure as Code with Terraform.
• Familiarity with ML lifecycle platforms such as MLflow, Kubeflow, Weights & Biases, or SageMaker Pipelines.
• Solid software engineering practices: unit/integration testing and secure coding.

Bonus

• Experience deploying and optimizing LLMs, including tools like vLLM and Hugging Face TGI, or optimization of open-source models (e.g., Llama, Mistral).