Forward Deployed AI Research Scientist

AMD AMD · Semiconductors · Santa Clara, CA · Engineering

AMD is seeking a Forward Deployed AI Research Scientist to integrate advanced AI capabilities into engineering and customer-facing workflows. This role involves applied AI research, technical strategy, prototyping, and collaboration with internal teams and external partners to identify problems, shape AI efforts, and drive measurable outcomes. The position bridges technical and business aspects, translating needs into research directions and AI capabilities into engineering value.

What you'd actually do

  1. Work with internal and external stakeholders to identify high-value AI opportunities in compute, systems, software, and hardware engineering workflows.
  2. Translate partner and customer needs into research questions, technical plans, evaluation criteria, and prototype paths.
  3. Prototype and evaluate modern AI methods for engineering tasks, including agentic workflows, tool use, code generation, debugging, optimization, retrieval, and learning from feedback.
  4. Partner with research and engineering teams to define metrics, evals, benchmarks, and acceptance criteria for AI-assisted workflows.
  5. Build technical narratives, demos, and decision materials that help senior stakeholders understand tradeoffs, progress, risks, and strategic value.

Skills

Required

  • Strong background in AI, machine learning, systems, or applied research
  • Ability to turn ideas into prototypes and measurable experiments
  • Experience with modern generative AI methods, LLMs, agents, tool use, retrieval, post-training, evaluation, or applied ML systems
  • Strong programming ability in Python
  • Familiarity with ML frameworks such as PyTorch, JAX, TensorFlow, or similar
  • Ability to work with technical stakeholders across software, hardware, infrastructure, research, product, and business teams
  • Excellent communication skills, including the ability to explain complex AI and systems topics clearly to both technical and executive audiences

Nice to have

  • Experience in forward-deployed engineering, customer engineering, applied research, technical partnerships, solutions architecture, or strategic technical programs
  • Experience with GPU systems, CPU performance, compilers, distributed training/inference, hardware/software co-design, or performance engineering
  • Experience designing benchmarks, evals, experiment platforms, or metrics for AI systems
  • Publications, open-source contributions, technical blogs, demos, or shipped AI systems
  • Familiarity with semiconductor, datacenter, AI infrastructure, or high-performance computing environments

What the JD emphasized

  • modern generative AI methods
  • LLMs
  • agents
  • tool use
  • retrieval
  • post-training
  • evaluation
  • applied ML systems
  • correctness
  • performance
  • latency
  • reliability
  • explainability
  • agentic systems
  • tools
  • tests
  • profilers
  • simulators
  • validation systems
  • structured feedback
  • Evaluation design
  • objective graders
  • human-in-the-loop review
  • LLM-as-judge methods
  • code understanding
  • code generation
  • optimization
  • debugging
  • knowledge retrieval
  • workflow automation
  • Technical-business translation

Other signals

  • Applied AI research for engineering workflows
  • Agentic systems that use tools, tests, profilers, simulators, validation systems, and structured feedback
  • Evaluation design for ambiguous or high-value tasks
  • AI systems for code understanding, code generation, optimization, debugging, knowledge retrieval, and workflow automation
  • Technical-business translation for strategic customers, partners, engineering leaders, and cross-functional teams