Senior Machine Learning Engineer
Purpose of the role
This role has the same remit as the Principal Machine Learning Engineer — the engineering backbone of the agent estate — but at a contributor rather than owner level. The successful candidate will implement and help evolve the architecture set by the principal role, with a growth path toward taking on more ownership over time.
Key responsibilities
· Build and maintain components of the agent estate/foundry: deployment pipelines, monitoring, and reliability tooling for agentic and conventional ML systems.
· Implement CI/CD pipelines and infrastructure-as-code for ML and agentic services under architectural direction from the Principal MLE.
· Develop and deploy MCP servers and integrate A2A-based agent communication into existing services.
· Deploy conventional ML models to production and support their ongoing operation alongside newer agentic components.
· Contribute to greenfield build-outs that integrate with legacy systems and data sources, flagging integration risk early.
· Support production observability and evaluation practices (MLOps/LLMOps) — monitoring, logging, alerting — for models and agents in the estate.
· Work day-to-day alongside the data science group (Senior Data Scientist / Senior AI Engineer) to turn model and agent designs into deployed, reliable services.
Essential experience
· 4+ years in machine learning or platform engineering.
· Experience building AI infrastructure on Azure (preferred) or AWS.
· Direct experience deploying ML models to production (not just training/experimentation).
· Some exposure to greenfield development integrated with legacy systems.
· Familiarity with the practical challenges of CI/CD in an ML context.
Essential skills
· Python.
· Docker and/or Podman.
· Terraform and CI/CD.
· SQL.
· MCP server deployment.
· A2A (Agent2Agent protocol) — hard requirement.
· Working knowledge of MLOps and LLMOps.
Desirable / nice-to-have
· Azure AI Foundry exposure.
· RAG systems and vector stores (pgvector, Pinecone).
· ETL and DAG orchestration.
· Serverless architectures.
· Claude Agent SDK or another agent framework/SDK.
What a strong candidate looks like
A solid, hands-on platform/ML engineer who has shipped production ML systems and is comfortable picking up newer agent-specific protocols quickly. Likely 1–2 years off principal-level readiness, and looking for a role with a clear senior mentor and room to grow into architecture ownership.
Sourcing notes for the agency
· Target titles: Senior Machine Learning Engineer, ML Platform Engineer, MLOps Engineer, AI Infrastructure Engineer.
· As with the principal role, prioritise genuine agent-to-agent / tool-orchestration experience over exact A2A protocol name-matching, but confirm conceptual understanding at interview.
· Good candidates may currently sit in a “conventional MLOps” role and be actively looking to move into agentic/LLM infrastructure — worth targeting explicitly.
Screening flags / watch-outs
· Same GCP-only caveat as the principal role — a GCP-only background is not workable given ramp-up time.
