Staff Agentic AI / Data Engineer

What you'll do

• Build production agentic AI systems on AMD Instinct GPU infrastructure including orchestration and tool calling.
• Design and operate memory and context data layers using vector, graph, and relational stores.
• Develop pipelines to ingest deployment telemetry and field knowledge into queryable formats for internal use.
• Maintain a library of reusable, versioned, and tested agent-executable skills and tools.
• Build evaluation infrastructure including regression suites and LLM-as-judge pipelines to monitor behavior.
• Harden agentic systems against security risks like prompt injection and data poisoning.
• Create reference architectures and open artifacts to establish AMD as a credible platform for agentic workloads.
• Support customer-facing engineers by deploying and maintaining systems in live production environments.

What we're looking for

• Experience building production data pipelines and storage systems at scale (preferred).
• Hands-on experience building LLM-powered and agentic applications in production (preferred).
• Expertise in memory/context storage paradigms such as vector, graph, or hybrid retrieval architectures (preferred).
• Experience designing evaluation frameworks for non-deterministic systems (preferred).
• Proficiency in Python and modern data stack tooling including orchestration, streaming, and Kubernetes.
• Familiarity with GPU inference serving like vLLM or SGLang; ROCm/AMD Instinct experience is a plus (preferred).
• Open-source contribution history in the AI/ML or data infrastructure ecosystem (preferred).
• Bachelor's or Master's degree in Computer Science, Computer Engineering, Data Engineering, or equivalent practical experience (preferred).