AI Systems Engineer

At Postali, we believe the strongest teams will pair expert human judgment with intelligent systems. We have already begun building AI-assisted workflows across content, search, research, reporting analysis, and website publishing. The next step is to make those systems dependable, measurable and genuinely useful in day-to-day work. We are looking for an AI-native builder who uses AI as a problem-solving medium, not simply as a writing tool. If you naturally decompose messy work, design the right combination of agents, automation and human judgment, and keep improving the system after it reaches production, this may be the right role for you. About the RoleThe AI Systems Engineer owns the design, development, and day-to-day operation of Postali's agentic AI systems. This is a hands-on, individual-contributor role for someone who can move from an ambiguous business problem to a working, monitored system that people use. You will partner across content, SEO, client performance, paid media, development and operations. Your job is to understand how work is actually done, decide what should be handled or supported by AI and build the workflow that makes the result faster, more consistent or more capable. What You'll Do • Own Postali's portfolio of AI agents and AI-enabled workflows across content, SEO, research, reporting analysis, publishing support and internal operations. • Find high-value problems worth solving. Map the current process, identify the decisions and failure points, and choose the right combination of LLMs, deterministic logic, automation and human approval. • Design and maintain reusable agent skills, prompts, context files, knowledge and retrieval systems, structured outputs, tool connections and client-specific configurations. • Improve Postali's AI content system, including research, drafting, client context, factual and citation checks, brand and legal-quality controls, editorial handoffs and repeatable setup for new clients. • Create agentic workflows that help expert teams diagnose performance, analyze competitors, synthesize data, recommend next actions and turn approved decisions into executable work. • Engineer human-in-the-loop controls. Define what the system can do autonomously, what requires expert review, and what should never be automated, especially for client-facing claims, regulated content, sensitive data and production changes. • Connect agents to approved tools and data through APIs, webhooks, MCP, automation platforms and light production code. • Build evaluations before declaring a system ready. Measure accuracy, completeness, consistency, compliance, latency, cost and real-world usefulness, and use those results to improve or retire the workflow. • Own production reliability for AI systems, including logs, monitoring, alerts, versioning, permissions, model and token costs, fallback behavior, incident response and recovery. • Create reusable frameworks that can scale across clients while preserving the specific context, standards and judgment each law firm requires. • Document system architecture, models, data sources, prompts and skills, dependencies, decisions, credentials, runbooks and backup ownership. • Train teams to use the systems effectively, observe where adoption or trust breaks down, and improve both the technology and the operating process. • Stay current through disciplined experimentation. Test new models, tools and patterns against clear use cases and evidence rather than adopting technology for its own sake. Why This Role Is DifferentMany AI roles focus on strategy decks, isolated demos or tool recommendations. This role owns systems that colleagues depend on in real work. You will have room to experiment, but you will also be accountable for what happens after the prototype: evaluation, adoption, quality, reliability, cost, documentation and measurable business impact. The goal is not more AI activity. It is better systems and better work. What We're Looking For • A track record of personally building and operating AI, automation, software or data systems that other people use. Titles and exact years are flexible; evidence of thoughtful, production-minded work matters more. • You are genuinely AI-native. You can show how AI has changed the way you research, reason, build, debug and solve problems. • Hands-on experience with LLM-based agents, tool calling, structured outputs, retrieval or knowledge systems, context engineering, evaluations and human-approval workflows. • Clear communication and change leadership. You can explain an AI system to technical and nontechnical colleagues, write usable documentation, and help a team adopt a new way of working. • Practical integration ability with APIs, webhooks, authentication, JSON and automation platforms. You know when no-code is sufficient and when code is the more reliable choice. • Experience with MCP, vector or hybrid retrieval, orchestration frameworks, workflow automation platforms and observability for