AI Systems Engineer - Agent Orchestration and Memory

About Hyatus Living Hyatus Living operates furnished apartments and is developing an AI platform to coordinate guest services, housekeeping, maintenance, pricing, and internal operations. Our goal is to make apartment operations increasingly autonomous while maintaining service quality and operational control. The role We are seeking an AI Systems Engineer to design and deliver the orchestration, context, retrieval, and memory infrastructure that makes these agents effective. This is a hands on role with ownership from architecture through deployment, focused on systems that can execute complex tasks accurately and maintain continuity over time. Responsibilities • Design agent orchestration systems with task decomposition, model and tool routing, parallel execution, checkpoints, and reliable handoffs. • Develop context management strategies for extended workflows, including context assembly, compaction, and continuity across sessions and agents. • Build persistent memory systems that retain useful knowledge, preserve source attribution, and reconcile outdated or conflicting information. • Implement agentic search across documents, conversations, and operational databases using query planning, keyword and vector retrieval, reranking, and evidence grounded responses. • Connect agents to backend services and structured data through APIs, schema aware SQL, and permission controlled tools. • Establish evaluations and monitoring for task completion, retrieval quality, failure recovery, latency, and cost. Qualifications • Demonstrated experience building and operating AI agent workflows, retrieval systems, or related infrastructure. • Strong backend engineering skills in Python or TypeScript, with practical experience in SQL, databases, APIs, Linux, and cloud deployment. • Practical experience with embeddings, keyword and vector indexes, retrieval augmented generation, and persistent memory architecture. • Substantial experience with tools such as Claude Code, Codex, Cursor, MCP, or custom agent frameworks. • Strong technical judgment, clear communication, and the ability to validate results independently and take systems into production. Application Please include a relevant project, repository, or technical case study. Describe your contribution, the system architecture, how you managed context and memory, and the evidence you used to assess reliability. For confidential projects, an anonymized technical summary is welcome.