Applied AI Engineer, Agents

About The Role

Morena is leading the search for a Staff Applied AI Engineer to build and deploy production AI agents for complex enterprise workflows. This is a senior, hands-on engineering role for someone who can take difficult AI deployment problems from initial design through production and turn what is learned from individual implementations into reusable engineering patterns. You will work on AI agents operating across communication, operational, and transaction-heavy workflows in a large regulated industry.

These systems need to do substantially more than generate responses. They need to reason through multi-step processes, interact with tools and APIs, follow business rules, manage state, and operate reliably under real production constraints. This is not a traditional solutions engineering role. You will be expected to design, build, debug, evaluate, and ship production systems while working closely with customers, product engineers, and platform teams.

• Take technical ownership of sophisticated AI agent deployments from initial design through production
• Understand the underlying workflow, design the right agent architecture, integrate with customer systems, evaluate behavior, resolve edge cases, and ensure reliable production performance
• Own outcomes rather than simply completing one piece of the implementation

• Design and iterate on production AI systems involving agent orchestration, prompt and context design, tool use, structured workflows, external API integrations, state management, retrieval, business logic, human escalation, and failure recovery
• Apply strong engineering judgment about where probabilistic AI belongs and where deterministic application logic should take over

• Establish how agent quality is measured and improved using evaluation datasets, automated checks, model-based evaluation, regression tests, production monitoring, trace analysis, failure classification, and real-world outcome metrics
• Investigate why agents fail, identify the underlying cause, and improve the system rather than relying on repeated prompt adjustments

• Turn lessons from individual deployments into shared components, agent patterns, templates, internal tooling, integration approaches, evaluation methods, and deployment playbooks
• Make each difficult implementation faster and more reliable for future work

• Work closely with Product and Platform engineering to bring lessons from production deployments back into the core product
• Distinguish between customer-specific requirements, reusable platform capabilities, product gaps, integration problems, model limitations, and workflow design problems
• Influence the technical roadmap with real production experience

• Diagnose difficult agent behavior across models, prompts, context, integrations, tools, infrastructure, and customer systems
• Move comfortably between reading production traces, investigating failed tool calls, debugging APIs, reviewing prompt or context construction, analyzing evaluation results, tracking distributed-system failures, and writing production code

• Raise the engineering bar through technical reviews, architecture decisions, mentorship, and the quality of systems you personally build
• Provide guidance to other engineers while continuing to own significant production work

• 6+ years of professional software engineering experience with a strong record of building and operating production software
• Comfortable with Python, APIs and services, cloud infrastructure, databases, distributed systems, integrations, observability, testing, and production operations

• Hands-on experience building systems using modern LLMs, including agentic systems, tool-calling, prompt and context engineering, LLM workflows, retrieval, structured generation, model APIs, and agent frameworks
• Production experience is significantly more important than experimentation alone

• Approach AI systems as production software with reliability, failure modes, testing, observability, data flow, API design, deployment, and operational risk in mind
• Identify whether issues come from the model, surrounding software, an integration, available context, workflow design, or evaluation method

• Reason systematically about…