Lead MLOps Platform Architect

Jobs Paloaltonetworks; United Kingdom; On-site; GBP 120,000 - 180,000; Full time

Job summary

Palo Alto Networks is seeking a Principal MLOps Engineer to join the Data & AI group in Cortex Research. You will design and scale the MLOps and LLMOps platforms powering data scientists and security researchers, building high‑performance infrastructure to train, fine‑tune, and deploy advanced AI systems across SLMs and RAG workflows.

You will own the serving architecture for LLMs/SLMs, optimize GPU utilization, and implement robust monitoring and automated training loops.

Qualifications

• 4+ years as Senior ML Engineer, MLOps Engineer, or Backend Platform Engineer (Hands‑On) in cloud environments.
• Experience managing lifecycle of diverse model architectures (ML, LLMs/SLMs, agentic/RAG) and scalable data pipelines.
• Experience with distributed training across multi-GPU setups using PyTorch/DeepSpeed/Megatron-LM or cloud-native infra.
• Strong cloud infra knowledge in GCP/AWS/Azure and managed AI platforms.

Responsibilities

• Scale distributed training with multi-GPU workloads and optimized compute.
• Automate the ML lifecycle with end-to-end pipelines for CT/CD.
• Own serving architecture for LLMs/SLMs, balancing latency and throughput.
• Implement comprehensive observability for live model performance and data drift.
• Collaborate with data scientists and security researchers; integrate with core cloud infra.

Skills

Distributed Training

ML Platform

Python

CI/CD

Cloud Platforms

GPU Compute

Observability

Collaboration

Tools

PyTorch

DeepSpeed

Megatron-LM

Kubernetes

GitHub Actions

GitLab CI

Job description

Palo Alto Networks is seeking a Principal MLOps Engineer to join the Data & AI group in Cortex Research. You will design and scale the MLOps and LLMOps platforms powering data scientists and security researchers, building high‑performance infrastructure to train, fine‑tune, and deploy advanced AI systems across SLMs and RAG workflows.

You will own the serving architecture for LLMs/SLMs, optimize GPU utilization, and implement robust monitoring and automated training loops.