Principal AI Systems Engineer- Agentic and Productivity Systems

The Opportunity

The Creative Cloud Engineering organization is building the next generation of AI-powered engineering infrastructure to accelerate developer productivity and operational excellence across the Creative Cloud ecosystem. As we expand into AI-driven workflows across developer productivity and platform initiatives, we are looking for a Senior AI Systems Engineer who operates at the intersection of experimentation and production systems. This role focuses on designing, orchestrating, and operationalizing agent-based systems that improve engineering workflows across CI/CD, developer tooling, and operational diagnostics.

This is not a research role and not a prompt-engineering role. This is a systems engineering role focused on building durable infrastructure. You will help build AI-native engineering capabilities that compound engineering velocity across Creative Cloud over time.

What You'll Do

Agentic Workflow Development

• Design and prototype agent-based systems for engineering workflows such as CI diagnostics, code review automation, build failure triage, and autonomous debugging

• Develop multi-agent orchestration patterns with structured state, memory, and control boundaries

• Rapidly evaluate emerging AI frameworks, agent tooling, and developer AI platforms in real-world engineering environments

AI Systems Infrastructure

• Build reusable orchestration layers and service architectures for AI-powered engineering systems

• Develop structured evaluation pipelines including trace-based evaluation and regression testing for agent behavior

• Implement feedback loops and instrumentation that continuously improve AI system performance

Production Hardening

• Convert experimental workflows into secure, scalable, production-grade services

• Implement observability, tracing, cost controls, and model routing

• Ensure reliability, operational stability, and measurable impact of AI-powered systems

Platform Strategy & Collaboration

• Define internal standards for AI experimentation, evaluation, deployment, and monitoring

• Partner with Dev Ex, CI/CD, and platform teams across Creative Cloud to embed AI-native capabilities

• Build cohesive infrastructure that prevents tool sprawl and enables reusable AI productivity systems across teams

What Success Looks Like

• Production-grade AI agents integrated into engineering workflows and CI systems

• A standardized evaluation and tracing framework adopted across Creative Cloud engineering teams

• Measurable reductions in manual debugging, failure triage, and operational friction

• Reusable AI infrastructure components leveraged across multiple engineering teams

• A clear AI productivity roadmap aligned with Creative Cloud platform initiatives

Required Qualifications

• 8+ years of software engineering experience, with demonstrated depth in systems-level work

• Strong systems engineering experience (Python, Go, Type Script, or similar)

• Experience building distributed systems, developer platforms, or infrastructure services

• Experience integrating LLMs or AI APIs into production systems

• Experience evaluating and integrating across multiple AI providers (e.g., AWS Bedrock, Anthropic, OpenAI) including cost optimization and capacity planning

• Strong understanding of observability, metrics, logging, and tracing systems

• Experience operating production services at scale

Preferred Qualifications

• Experience with agent frameworks (Lang Graph, Auto Gen, CrewAI, or similar)

• Experience with embeddings, vector databases, or RAG architectures

• Experience designing evaluation and benchmarking systems for AI workflows

• Experience with CI/CD platforms, developer tooling, or build systems

• Experience building internal developer productivity platforms

• Familiarity with cost-aware model orchestration and multi-model routing

Ideal Candidate Profile

• Has built and shipped an AI-powered system end-to-end, not just integrated an API

• Can show a prototype they took from experiment to production

• Comfortable making infrastructure decisions with incomplete information

• Has debugged LLM reliability issues in production (latency, cost, failure modes, concurrency limits)

• Experimental but pragmatic -…