Job Title: DevOps Engineer with AI/ML and MLOps Location: New Jersey (Day 1 Onsite) Duration: Contract Job Description : • Document architecture workflows and best practices for knowledge sharing and compliance. • Provide technical oversight & Guidelines. • Architect and implement end-to-end MLOps and LLMOps pipelines using Azure Machine Learning and Azure OpenAl. • Design scalable infrastructure for training, deploying, and monitoring ML and LLM models in production. • Collaborate with data scientists and engineers to streamline model development testing, and deployment workflows. • Manage Azure Kubernetes Service (AKS) clusters and containerized ML workloads. • Ensure model governance, versioning, and reproducibility using tools like MLflow and Azure DevOps • Promote DevSecOps practices, ensuring security and compliance are embedded in the ML lifecycle. • Monitor and troubleshoot production ML systems, ensuring high availability and performance. • Experience with Azure Machine Learning. Azure OpenAl, Azure DevOps, and AKS. • Proficiency in Python, Docker, Kubernetes, and CI/CD pipelines. • Experience with LLM fine-tuning, prompt engineering, and model deployment. • Familiarity with MLflow, Terraform, and monitoring tools like Prometheus/Grafana. Responsibilities • Design and maintain CI/CD pipelines tailored for Al worldlows (eg. RAG, chunking, deployment). • Familiarity with DABs or other infrastructure-as-code deployment patterns. • Version control for prompts using MLFlow. • Automate prompt testing pipelines to evaluate performance across different LLMs or datasets using evaluation frameworks and MLFlow. • Support A/B testing of prompts in production environments. • Automate ingestion pipelines. • Containerize and orchestrate RAG components (retriever, ranker, generator). • Integrate retriever and generator components into a unified API endpoint. • Optimize latency and throughput for real-time RAG queries. • Experience with ingestion of unstructured documentation and OCR technologies Experience with OCR optimization/parallelization in production.