Principal Architect, Express AI Foundations

Adobe Adobe · Enterprise · San Jose, CA +1

Principal Architect to build and implement the AI framework for Adobe Express, merging strong ML skills with proficiency in distributed systems, data architecture, and large-scale service development. This role blends applied research and engineering leadership, focusing on Agentic AI, Create AI, Imaging AI, Motion AI, and Personalization AI, spanning model orchestration, inference systems, data pipelines, caching and storage layers, session analytics, and continuous evaluation frameworks.

What you'd actually do

  1. Architect and evolve the complete AI stack for Adobe Express — covering Agentic AI, Construct AI, Imaging AI, Motion AI, and Personalization AI.
  2. Develop and operationalize end-to-end systems — integrating microservices, data pipelines, LLM orchestration layers, in-house and third-party models, databases, caches, session analytics, and evaluation systems into a cohesive architecture.
  3. Develop large-scale data and inference infrastructure to support model training, fine-tuning, evaluation, and deployment — employing Spark, Kafka, Flink, and other distributed frameworks.
  4. Develop high-performance runtime services for inference and orchestration with strong observability, fault tolerance, and latency guarantees.
  5. Lead development of experimentation and evaluation systems, encompassing session-level analytics, feedback loops, and quality metrics that drive continuous improvement.

Skills

Required

  • Python
  • Java
  • C++
  • Go
  • distributed systems
  • cloud-native deployment
  • performance tuning
  • ML fundamentals
  • LLM fundamentals
  • training
  • fine-tuning
  • deployment
  • evaluation workflows
  • data pipelines
  • real-time streaming systems
  • event-driven architectures
  • caching strategies
  • database development
  • large-scale serving systems
  • LLM orchestration frameworks
  • model routing
  • multi-model inference
  • Agentic AI patterns
  • reasoning loops
  • memory persistence
  • task decomposition
  • multi-agent coordination

Nice to have

  • MLOps pipelines
  • feature stores
  • model registries
  • open-source contributions
  • publications
  • conference presentations
  • AI assistants
  • build agents
  • multimodal creative systems
  • Generative AI
  • LLMs
  • diffusion models
  • multimodal architectures

What the JD emphasized

  • 10+ years of experience in large-scale distributed systems AI infrastructure, or ML platform engineering.
  • Deep understanding of ML and LLM fundamentals — training, fine-tuning, deployment, and evaluation workflows.
  • Proven expertise in building and scaling data pipelines, real-time streaming systems, and event-driven architectures (Kafka, Spark, Flink, etc.).
  • Strong background in caching strategies, database development, and performance optimization for large-scale serving systems.
  • Hands-on experience with LLM orchestration frameworks, model routing, and multi-model inference.
  • Familiarity with Agentic AI patterns — reasoning loops, memory persistence, task decomposition, and multi-agent coordination.

Other signals

  • building and scaling data pipelines
  • real-time streaming systems
  • event-driven architectures
  • LLM orchestration frameworks
  • multi-model inference
  • Agentic AI patterns