Senior Software Engineer (AI-Native)

Location: Office located near Chatham, NJ (Northern NJ) — full-time, in-person

Reports to: Founder / CEO

Base salary: $135,000–$175,000

Before you apply: You must already be authorized to work in the United States. We cannot sponsor or take over sponsorship of an employment visa. We are not considering remote candidates, contractors, agencies, or offshore development firms.

About Avalon Tax

Avalon Tax builds intelligent tax preparation automation software for tax professionals (including CPAs and EAs). Our desktop application reads client tax documents, organizes the return workpapers, transfers data into professional tax software, and gives the preparer clear review points before anything moves forward. Today, Avalon Tax works with ProConnect, Drake, Lacerte, and CCH Axcess Tax. The preparer remains in control. Accuracy, explainability, secure handling of taxpayer information, and human review are central to how we build.

The Role

Avalon Tax is a shipped product with real customers and revenue. We are hiring a senior software engineer to own the health of that product: keep it fast, reliable, and secure as our customer base grows, and ship new features alongside our Senior Tax Product Specialist, who defines what a correct tax result looks like while you make the software deliver it.

​

You will be the most experienced engineer on the team, but not the only person in the code. The founder, the tax product specialist, and AI coding agents all work in the repository daily. Part of your job is keeping quality high across all of it: reviewing changes whether a person or an agent wrote them, strengthening tests and CI, and improving the tooling the whole team relies on.

​

How we work matters as much as what we build. AI coding tools are load-bearing in our development process; we use them to move at a speed a team our size could not otherwise reach, with review and verification practices to match. We have learned firsthand that engineers who prefer to work without these tools slow the whole team down. You should be someone who already gets real leverage from AI agents every day, reviews their output critically rather than accepting it on faith, and enjoys making the tools themselves work better.

What You’ll Do

• Own the reliability, performance, and security of the product across the full stack: an Electron + React (TypeScript) desktop app, a Python FastAPI document extraction backend, and a Node.js/Express + PostgreSQL cloud backend.
• Scale the platform as we add customers: throughput and resilience of the document extraction pipeline, observability, error handling, and cost control of third-party AI services.
• Keep taxpayer data safe: local-first architecture, privacy-preserving telemetry, authentication, secrets handling, dependency hygiene, and secure defaults everywhere.
• Partner with the Senior Tax Product Specialist to turn tax requirements into shipped features, from document extraction through automated data entry into ProConnect, Drake, Lacerte, and CCH Axcess Tax.
• Troubleshoot customer-reported issues end-to-end across desktop and cloud, fix root causes rather than symptoms, and turn every fix into a regression test.
• Review and integrate code from humans and AI agents alike; raise the bar on testing, CI, and release quality.
• Improve our AI-assisted development infrastructure (e.g agent workflows, validation hooks, review tooling) so the entire team ships faster and safer.
• Own our third-party integrations across AI and document-intelligence services, billing, identity, and CRM tooling.

Required Qualifications

• Self-driven problem solver who enjoys a challenge and is naturally curious how we can use AI to improve the working lives of our customers
• 7+ years of professional software engineering experience, including end-to-end ownership of production systems that real customers depend on.
• Strong in at least two of our core technologies (Python, TypeScript/React, Node.js) and able to work effectively across all of them. Experience with relational databases (we use PostgreSQL) and REST APIs.
• Demonstrated AI-native development: AI coding agents (Claude Code, Codex, Cursor, or comparable) are already part of how you work daily; you direct them with the right context and constraints, inspect and test what they produce, and iterate to a verified result. Be prepared to show real examples.
• A security mindset: we handle sensitive taxpayer information, and you treat that as a design cons…