Senior Agentic Systems Engineer

About the Company

Our client is building an AI-powered marketing SaaS for law firms and professional services teams, automating content generation, research, and multichannel publishing in a domain where quality, accuracy, and compliance matter. The product sits at the intersection of agentic AI systems, workflow automation, and legal-content governance, with a growing platform that now needs stronger backend foundations for reliability, tenant isolation, and operational visibility.

About the Role

This is a backend ownership role with meaningful architectural scope. You’ll lead the evolution from a fragmented ECS/N8N setup to a cleaner Next.js monolith with native async workers, while designing the AI pipelines, job systems, and multi-tenant data model that support the product end to end. It’s a strong fit for an engineer who has built production AI workflows, cares about observability and failure handling, and can translate complex requirements, like legal-content compliance, publishing reliability, and billing accuracy, into durable systems.

What You’ll Build Migrate a Node.js/Express ECS backend into a Next.js monolith with API routes Replace N8N with native BullMQ async workers for all content generation pipelines Implement AI observability (Langfuse or LangSmith) across Claude, Perplexity, Gemini, and

DALL-E call sites Design and enforce Supabase RLS policies for strict tenant isolation Build dead-letter queues, retry strategies, and alerting for async AI job failures Implement Stripe Checkout, tier gating, webhooks, and customer portal end-to-end Architect the compliance layer: CourtListener API integration, bar-rule checks by state, audit trail per published piece Add per-client token cost tracking across all AI providers Instrument WordPress REST API publishing with retry logic and per-client health monitoring

Required Experience 5+ years full-stack engineering; 2+ years building multi-tenant SaaS Next.js (App Router, API routes, server components) TypeScript, Node.js, PostgreSQL / Supabase BullMQ or equivalent async job queue architecture in production

• Supabase RLS: designed and enforced tenant isolation, not just read the docs
• Agentic AI systems: prompt chaining, fallback logic, output validation, non-deterministic testing AI observability tooling (Langfuse, LangSmith, or Helicone) Stripe billing end-to-end: Checkout, webhooks, customer portal, metered billing OAuth integrations (Google, LinkedIn)

Aws (Ec2, S3, Ecs Familiarity)

Nice to Have

• Experience with Claude API tool use / extended thinking
• Compliance or legal tech background
• Prisma ORM

Not a Fit If

• You treat AI API calls as just another fetch request
• You’ve only built single-tenant apps
• Observability and evals aren’t part of your default workflow

Engagement: Full-time contract, remote