Software Engineer, Agentic AI Infrastructure

Socket.dev; San Francisco (CA); On-site; USD 120,000 - 160,000; Full time

Job summary

Anrok, located in San Francisco, is seeking an engineer to build foundational AI agent infrastructure that operates reliably at scale. You'll leverage your systems engineering background and experience with AI agents to ensure scalable and fault-tolerant services.

This role involves designing and shipping the systems that empower AI agents to manage tax data efficiently. The ideal candidate will have hands-on experience with distributed systems and is eager to collaborate with product engineers to translate compliance workflows into robust solutions.

Qualifications

• 3+ years of experience building and operating distributed, production-grade systems.
• Experience with AI agents deployed in production, focusing on reliability and maintenance.
• Strong understanding of scalable architecture and tradeoffs in technical decisions.

Responsibilities

• Design and ship foundational infrastructure for AI agents at Anrok.
• Own end-to-end technical architecture decisions for agent infrastructure services.
• Collaborate with product engineers and tax data teams to create scalable agent solutions.

Skills

Systems engineering

Distributed systems

AI agent systems

Architecture decision-making

Software development lifecycle

Prompt engineering

Tools

AI stack

Deployment tools

Job description

Requirements

• You have a strong systems engineering background, with 3+ years of experience building and operating distributed, production-grade systems
• You've designed services that are scalable, resilient, and fault-tolerant—and you've done it more than once
• You've built and deployed AI agents or LLM-powered systems in production, and you understand firsthand what makes them reliable and maintainable at scale
• You're familiar with the modern AI stack—prompt engineering, eval frameworks, RAG pipelines, tool calling, and agent orchestration—and you have a point of view on what the infrastructure beneath it should look like
• You think carefully about architecture and tradeoffs, and you're comfortable leading technical decisions without a complete playbook
• You have experience with the full software development lifecycle: design, code review, testing, deployment, and on-call ownership
• You're energized by ambiguity—you can take a fuzzy problem, break it down, and ship something real
• You're curious and motivated to learn new tools—you've experimented with AI in your work or on your own, you're excited about what's possible, and you've built something with it (an application, a workflow, a creative solution to a real problem) that you can walk us through

What the job involves

• We're building the agentic infrastructure layer at Anrok—the systems that let AI agents reason about tax data, take action across jurisdictions, and operate reliably at scale
• This is early and foundational work: you'll be designing and shipping the primitives that the rest of the product builds on. We're looking for an engineer who pairs a strong distributed-systems background with real, hands‑on experience building AI agent systems—someone who understands both how to make a service fault‑tolerant and what it takes to make an LLM‑powered agent reliable in production
• If you want to work on distributed systems that sit at the intersection of AI and fintech, this role is for you
• Design and build the distributed, fault‑tolerant infrastructure that powers Anrok's AI agent capabilities—covering memory, state management, and execution for autonomous systems
• Own technical architecture decisions for new agent infrastructure services end‑to‑end, from initial design through production deployment and iteration
• Build the evaluation, observability, and reliability tooling that lets the rest of the team ship agents confidently on top of your infrastructure
• Collaborate closely with product engineers and our tax data team to translate real compliance workflows into reliable, scalable agent primitives
• Drive engineering best practices across the team: code quality, testing discipline, system design, and operational health
• Identify and fix performance, reliability, and scalability bottlenecks as our agent systems grow
• Help grow the team—mentoring engineers, contributing to technical culture, and raising the bar on how we build