AI / ML Ops Engineer

Role overview

Responsibilities

• Design and automate end-to-end ML pipelines for continuous training and deployment using orchestration engines.
• Deploy containerized models with low-latency inference, autoscaling GPU/CPU clusters, and Kubernetes serving runtimes.
• Manage feature stores, model registries, and dataset version control foundations.
• Automate monitoring of model accuracy decay, data and concept drift, and system latency.
• Optimize inference using model compilation and quantization to reduce latency and compute costs.
• Operate generative AI and LLM systems, including semantic caching, vector database scaling, and prompt validation.
• Configure ML access controls, data separation, access tokens, and encryption to protect sensitive inference logs.

Required qualifications

• 4–8 years of core software engineering, DevOps, or data engineering experience, including 3+ dedicated years building and maintaining MLOps automation infrastructure.
• Strong technical mastery of Python, Docker, Kubernetes, PyTorch, TensorFlow, Hugging Face, and advanced SQL.
• Deep understanding of distributed systems, GPU resource management, deployment patterns such as Shadow, Canary, and A/B, and cloud provider API governance.
• One certification is mandatory: AWS Machine Learning Specialty, Google Cloud Professional ML Engineer, or Databricks ML Professional.

Tools and platforms

• Pipeline orchestration examples include Kubeflow, MLflow, and AWS SageMaker Pipelines.
• Model serving examples include KServe and Triton Inference Server; feature-store example is Feast; DVC is cited for dataset versioning.
• Inference optimization examples include ONNX and TensorRT; vector database examples include Pinecone and Milvus.

Preferred qualifications

Work arrangement

• The position is remote; the job details also state offshore location.

Employment details