Software Engineer, Agentic AI Infrastructure

About this role

Anrok seeks a Software Engineer to build foundational agentic AI infrastructure that powers tax automation at scale. You'll design distributed, fault-tolerant systems for AI agents to reason about tax compliance and operate reliably across multiple jurisdictions, combining deep systems engineering expertise with hands-on production AI experience.

What you'll do

• Design and build distributed, fault-tolerant infrastructure for AI agent memory, state management, and execution
• Own end-to-end technical architecture for new agent services from design through production deployment
• Develop evaluation, observability, and reliability tooling for confident agent deployment
• Collaborate with product and tax teams to translate compliance workflows into scalable agent primitives
• Drive engineering best practices around code quality, testing, and system design
• Identify and fix performance, reliability, and scalability bottlenecks in agent systems

What they're looking for

• Distributed systems design and implementation (3+ years)
• Production AI agent and LLM system development
• Scalable, resilient, fault-tolerant service architecture
• Modern AI stack (prompt engineering, eval frameworks, RAG, tool calling, agent orchestration)
• Full software development lifecycle (design, testing, deployment, on-call)
• Systems thinking and technical decision-making
• Hands-on experience building AI applications or workflows

Benefits

• Equity upside with early-stage startup that has product-market fit
• 100% covered medical, dental, and vision insurance
• One Medical membership covered
• Daily lunch and snacks at San Francisco office
• Annual learning and development stipend
• Team offsites and in-person opportunities