MLOps Engineer

Job description

**Skills Required:** - MLOps - LLMOps - Python - SQL - ML - AI - RAG **Roles & Responsibilities:** • Build and manage MLOps and LLMOps pipelines. • Deploy, host, and scale Deep Learning models, LLMs, and SLMs and Inference optimisation • Manage end-to-end model lifecycle including versioning, deployment, rollout, rollback, and retirement. • Host models on Databricks, Kubernetes, OpenShift, and GPU-based infrastructure. • Implement model governance, lineage, approval workflows, and compliance controls. • Build model monitoring, observability, tracing, logging, and drift detection capabilities. • Optimize model performance, latency, throughput, GPU utilization, and cost. • Support cloud, on-premises, hybrid, and air-gapped environments. **Must-Have Skills:** • 35 years in MLOps, LLMOps, ML Engineering, or AI Engineering. • Strong Python development skills. • Hands-on experience with Databricks and/or Azure ML. • Experience with Deep Learning, LLMs, SLMs, RAG, and Hugging Face. • Experience deploying models built using PyTorch and TensorFlow. • Strong expertise in model deployment on: o Kubernetes o Databricks o GPU Infrastructure **Experience with:** o LLM o Triton Inference Server o Ray Serve o SGLang o Databricks Model Serving • Strong GPU knowledge including NVIDIA GPUs, CUDA, multi-GPU deployments, and inference optimization. • Experience in Model Registry, Model Governance, Model Monitoring, Drift Detection, and AI Observability. • Strong database knowledge (SQL Server, PostgreSQL, Oracle, MySQL, MongoDB). • Experience with Vector Databases (Pinecone, Chroma, FAISS, Milvus, Azure AI Search). • REST APIs, WebSocket, Streaming HTTP. • Experience with MLflow, Open Telemetry, Lang Fuse, Splunk, and Grafana/ELK. • CI/CD using Jenkins, Azure DevOps. • Experience across Cloud, On-Premises, Hybrid, and Air-Gapped environments. • Experience with Auth setup like Key cloak **Good-to-Have Skills** • Kafka, RabbitMQ, Event Hub • Fine-tuning and model optimization • Model Governance & Security • Experience with Llama, Mistral, DeepSeek, Qwen, Phi, and Gemma models