Software Engineer, AI

About Clark

Every audit answers one question: can I trust this? For 150 years, people have answered it by hand. Clark is changing that.

We build AI that doesn’t just assist auditors. It does the audit work itself. More than 35 CPA firms already use Clark to run audit fieldwork faster and at equal or better quality. Audit is where we start. Our mission is to build the trust infrastructure layer for the AI economy, and we’re only getting started.

We’re a small, hard-working team backed by top investors. We take ownership and judge ourselves by results. If you want to change how an entire profession works, you’ll fit in here.

The Role

We are hiring Senior Engineers to grow the Ava team and build the AI systems that power Clark's core product.

You will design and ship production AI systems that automate complex audit and accounting workflows. This includes LLM-powered agents, retrieval systems, document intelligence, evaluations, validation, and the infrastructure required to make AI reliable in high-trust environments.

You will work closely with product engineers, Audit Engineers, customers, and company leadership to identify high-value workflows and build systems that can perform them accurately, securely, and at scale.

Why This Role Matters

Clark's ability to win and retain customers depends on the depth, reliability, and quality of Ava.

As customers ask Clark to automate more complex workflows, the Ava team must build AI systems that are both impressive in demos and reliable in production.

This role directly expands what the business can sell, deploy, and support.

What You'll Do

• Build and improve LLM-powered workflows within Ava

• Design agentic systems for audit and accounting workflows

• Improve retrieval, reasoning, evaluation, and reliability of AI outputs

• Partner with Audit Engineers to understand real customer workflows

• Create evaluations, monitoring, guardrails, and human-in-the-loop review systems

• Build production systems that can operate in trust-sensitive environments

• Improve prompt, tool-use, orchestration, and evaluation systems

• Improve the accuracy, reliability, latency, and cost of production AI workflows

• Collaborate with platform engineers on reliability and observability

• Debug AI failure modes and improve system performance over time

• Ship product features that directly impact customer adoption

• Debug failures in live customer workflows and improve system performance over time

• Help establish Clark’s technical standards for building safe and dependable AI systems

Who You Are

• 2-5 years of software engineering experience

• Experience building production software systems

• Strong backend or full-stack engineering fundamentals

• Hands-on experience building applications with LLMs, agents, retrieval systems, or other modern AI models

• Experience designing evaluations and improving AI-system performance using real production data

• Comfortable working with probabilistic systems and designing around uncertainty and failure

• Strong product judgment and an ability to identify the simplest system that delivers meaningful customer value

• Comfortable working directly with domain experts and translating complex workflows into software

• High ownership, strong communication, and excellent technical judgment

• Excited to operate in person with a fast-moving, talent-dense team

Bonus

• Experience building production LLM applications

• Experience with agent frameworks, evals, retrieval systems, or tool-calling

• Experience in fintech, compliance, accounting, audit, legaltech, or other trust-sensitive software

• Experience with document-heavy workflows

• Experience at a fast-moving startup

The Environment

• 5 days a week on-site in NYC — no remote or hybrid

• Startup hours and pace — this isn't a 9-to-5

• Flat structure, no hierarchy — everyone ships

• High ownership — you own outcomes, not just tasks

Benefits

• Unlimited PTO

• Wellhub Platinum membership

• Comprehensive health, dental, and vision coverage

• Free company dinner every day