AI Software Architect, Forward Deployed

Adobe Adobe · Enterprise · London, United Kingdom

This role is for an AI Software Architect focused on Forward Deployed Engineering for enterprise customers. The individual will own system design, platform standards, and technical relationships with large enterprise clients, transforming their content supply chains with generative AI. Responsibilities include designing end-to-end GenAI systems (RAG pipelines, agent orchestration, API integrations), establishing reusable patterns, shipping production code (minimum 40% hands-on), interfacing with C-level customers, driving product feedback, and mentoring engineers. The role requires experience building and shipping production GenAI products, a deep AI/ML engineering foundation, strong production engineering rigor, and executive communication skills.

What you'd actually do

  1. Own system architecture end-to-end.
  2. Set platform standards.
  3. Ship production code.
  4. Interface with C-level customers.
  5. Drive the product feedback loop.

Skills

Required

  • system design
  • platform standards
  • technical relationship with enterprise customers
  • production code
  • architecture
  • GenAI systems
  • RAG pipelines
  • agent orchestration
  • production-grade API integrations
  • reusable patterns and frameworks
  • product engineering
  • transformer architectures
  • diffusion models
  • embeddings
  • vector databases
  • fine-tuning
  • prompt engineering
  • full-stack development (React/Next.js, Node.js/Python)
  • PostgreSQL
  • distributed systems
  • CI/CD
  • observability
  • error budgets
  • scalable infrastructure (Docker, Kubernetes, cloud-native)
  • AI tools (Cursor, Claude Code, Copilot)
  • MCP servers
  • custom agents
  • agentic workflows
  • 8+ years of shipping production software
  • 3+ years building AI/ML systems
  • executive communication
  • startup speed

Nice to have

  • contributed to open-source AI projects
  • built tools others use
  • published technical writing
  • experience with creative/content production workflows
  • built multi-agent systems
  • complex agentic architectures in production

What the JD emphasized

  • production code
  • built and shipped GenAI products in production
  • Not demos. Not POCs that died. Real systems handling real traffic
  • You build the thing AND design the system.
  • Minimum 40% of your time is hands-on.
  • AI-native builder
  • built MCP servers, custom agents, or agentic workflows
  • AI tools are how you work, not something you've read about.

Other signals

  • design end-to-end GenAI systems
  • RAG pipelines
  • agent orchestration
  • production-grade API integrations
  • reusable patterns and frameworks
  • ship production code
  • build and shipped GenAI products in production
  • AI/ML engineering foundation
  • transformer architectures
  • diffusion models
  • embeddings
  • vector databases
  • fine-tuning
  • prompt engineering
  • AI-native builder
  • built MCP servers, custom agents, or agentic workflows