Data Scientist – Agentic AI

About the role

The Data Scientist – Agentic AI will design, implement, and operationalize core agentic workflows that power Marvis, HPE's next‑generation AI assistant for network operations. This hybrid role requires working on‑site two days per week.

Key responsibilities

• Design and iterate agentic workflows using LangGraph, including ReACT loops, dynamic tool selection, multi‑step reasoning and self‑correction.
• Develop and maintain Model Context Protocol (MCP) servers, define tool schemas, implement domain‑specific tools and manage server lifecycle.
• Integrate LLM capabilities at production scale, handling structured outputs, streaming, function calling, prompt engineering and robust error handling.
• Build retrieval‑augmented generation (RAG) pipelines, vector‑store semantic search and long‑term agent memory services.
• Create evaluation frameworks for non‑deterministic systems, define metrics, run A/B tests and drive continuous quality improvement.
• Collaborate with network engineers and product managers to translate troubleshooting playbooks into skills, tools and data pipelines.
• Deploy and operate containerised services in Kubernetes, contribute to CI/CD pipelines and ensure observability through tracing, logging and performance monitoring.

Required profile

• Bachelor’s, Master’s or PhD in Computer Science, Data Science, Mathematics, Statistics or a related quantitative discipline.
• 4–7 years of experience building production ML/AI systems, with at least one year focused on generative AI and LLM‑based applications.
• Strong understanding of agentic design patterns such as ReACT loops, tool/function calling and self‑correction.

Required skills

• LangGraph (preferred), LangChain, Claude Agent SDK or equivalent agentic orchestration frameworks.
• Model Context Protocol (MCP) implementation and tool‑serving protocols.
• LLM API integration at scale, prompt engineering, structured outputs and cost optimisation.
• RAG pipeline design, vector stores (e.g., OpenSearch) and hybrid retrieval techniques.
• Python development in production codebases.
• Kubernetes deployment, CI/CD (Jenkins, GitHub Actions, ArgoCD), Docker container management.
• FastAPI or similar backend development, REST/GraphQL API design.
• Observability tools such as LangFuse, Prometheus, logging and tracing solutions.

What we offer

• Comprehensive health and wellbeing benefits.
• Personal and professional development programs.
• Inclusive workplace culture supporting flexibility and growth.