Senior / Prompt Engineer (ai/ml), Brand Concierge

Adobe Adobe · Enterprise · San Jose, CA

Senior Prompt Engineer at Adobe focused on enhancing LLM performance for enterprise use cases, designing multi-step agent workflows, implementing function calling and tool use, and optimizing prompts for accuracy, privacy, and compliance. The role involves collaboration with AI engineers and product teams to integrate prompts into production agent systems.

What you'd actually do

  1. Develop templates and multi-step chains tailored to specific business functions (e.g., sales enablement, support, knowledge management)
  2. Implement LLM function calling to trigger APIs, databases, or internal tools
  3. Define and build agent personas, roles, and behaviors for domain-specific applications
  4. Tailor prompts for performance in enterprise environments, prioritizing accuracy, privacy, fairness, and compliance
  5. Partner with AI Agent Engineers to integrate prompts into agent workflows and orchestration pipelines

Skills

Required

  • 3+ years of experience in NLP, AI/ML product development, or application scripting
  • Strong grasp of LLM capabilities and limitations (e.g., OpenAI, Claude, Mistral, Cohere)
  • Experience crafting prompts and evaluation methods for enterprise tasks
  • Strong ML, Python and API integration skills
  • Excellent written communication and structured thinking

Nice to have

  • Experience with LLM function calling, custom tool integration, and agent workflows
  • Background in UX writing, human-computer interaction, or instructional design
  • Understanding of enterprise compliance (e.g., SOC 2, GDPR) in AI systems
  • Familiarity with frameworks like LangChain, Semantic Kernel, or AutoGen

What the JD emphasized

  • Implement LLM function calling to trigger APIs, databases, or internal tools
  • Tailor prompts for performance in enterprise environments, prioritizing accuracy, privacy, fairness, and compliance
  • Collaborate with legal and security teams to mitigate hallucination, bias, and misuse risks
  • Strong grasp of LLM capabilities and limitations
  • Experience crafting prompts and evaluation methods for enterprise tasks

Other signals

  • LLM performance
  • intelligent agents
  • enterprise use cases
  • function calling
  • tool use
  • agent workflows