AI Engineer

We're looking for an AI Engineer to build and extend the AI-powered systems at the core of our product. You'll work across the stack, from designing LLM-driven agents that automate complex workflows, to building extraction and validation pipelines, to applying AI to new product features as we grow.

This is a hands-on engineering role. You'll own the systems you build, from prototype through production. You'll work closely with a small engineering team and have a direct impact on what we ship and how we ship it.

What you'll work on:

* AI-driven automation. Build agents and workflows that use LLMs and other models to automate tasks that traditionally require human judgment; interacting with external systems, evaluating rules, annotating data, and making decisions based on unstructured inputs.

* Document understanding and extraction. Improve and extend our AI pipelines that extract structured data from credential documents; working with vision models, prompt engineering, and structured output to turn uploaded files into validated, actionable data.

* Intelligent system integration. Build the AI layer that bridges our platform with external systems (ERPs, ATSs, VMSs and many more acronyms!) where data formats, terminology, and requirements vary. Translate between systems that don't speak the same language; mapping external credential requirements to internal document types, reconciling naming differences, and handling edge cases that pure rule-based approaches can't cover.

* Evaluation and quality. Design systems that use AI to evaluate and improve our own processes; validating rules, measuring extraction accuracy, identifying gaps, and generating training data through automated annotation.

* AI feature development. Apply LLMs, vision models, and other AI capabilities to new product features as they emerge. Evaluate new models and techniques, and figure out where AI adds real value vs. where simpler approaches win.

* Reliability at scale. Make AI-powered systems production-grade; retry logic, failure detection, confidence scoring, audit trails, and monitoring. We're scaling to handle high-volume customers with hundreds of credentialing analysts, so what you build needs to work reliably, not just impressively.

What we're looking for:

* 3-5 years of software engineering experience, with at least 1-2 years spent building with LLMs, vision models, or other AI/ML systems in production.

* Strong software engineering fundamentals. You write clean, maintainable code. You understand system design, async processing, and how to build things that don't break.

* Experience with LLM integration. You've built systems that use LLMs for real tasks; structured output, tool use, prompt engineering, working with multiple model providers.

* Proficiency in at least one modern, production grade language. Our primary stack is TypeScript, but we care more about strong engineering fundamentals than framework expertise. You should be able to pick up whatever tools the problem calls for.

* Experience designing evaluations for AI-powered features; measuring output quality, building feedback loops, and using data to iterate on model/pipeline performance

* Pragmatic problem-solving. You know when to use an LLM and when a regex is fine. You optimize for shipping, not for architectural purity.

Nice to have:

* Experience with vision models (document understanding, page/screenshot analysis)

* Experience building or working with AI agent frameworks

* Experience with browser automation

* Background in healthcare, staffing, or credentialing

* Experience with AWS (ECS, S3, Lambda)

* Experience with queue-based architectures and async job processing