AI Systems Engineer

Postali is a company focused on building AI-assisted workflows across various domains. They are seeking an AI Systems Engineer to design, develop, and manage AI systems that enhance operational efficiency and effectiveness.

Responsibilities:

• 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