Staff Software Engineer, Applied AI

Gravie builds health benefits for small and midsize businesses and their employees. The company was founded and is still led by healthcare industry veterans who have started and grown several market-leading companies in the space, and it is backed by top-tier investors.

How engineering works here — worth reading first

Gravie is unusually explicit about its engineering model, and it is the most distinctive thing about this role.

Engineers work closely with Product and own outcomes end to end: understanding the business deeply, shaping both the problem and the solution, and staying accountable for whether what ships actually solves the problem — not simply implementing a set of requirements.

They practise agentic development as standard. Engineers start from a clear outcome, use AI agents to help develop the specification and plan, then use those agents to execute substantial multi-step engineering work. The engineer reviews and refines the spec, sets guardrails and acceptance criteria, validates results, and owns the quality of everything shipped.

Their framing of what that means for the role is worth quoting in substance: as agents take on more of the execution, the bar for engineers rises. Problem framing, business understanding, architectural judgment, systems thinking, risk identification and the ability to evaluate quality and tradeoffs all become more important, not less.

What you'll do

• Design and build production-grade AI agent systems, including multi-agent workflows, orchestration layers, and the infrastructure to make them reliable, observable and auditable
• Design retrieval, context and memory architectures that ground AI outputs and support reliable multi-step workflows
• Lead AI-powered decision-support systems meeting the accuracy, traceability and explainability requirements of healthcare and other regulated environments
• Architect distributed services, event-driven workers, job pipelines, state management, failure recovery, APIs and data models
• Guide product-quality experiences in React and TypeScript that make complex AI capabilities understandable
• Establish organisation-wide patterns for AI quality and safety — automated evaluations, regression testing, groundedness checks, compliance guardrails, production monitoring
• Mentor senior engineers, lead architecture reviews and shape technical roadmaps

What you bring

• 10+ years of software engineering, including technical leadership of complex distributed systems from design through production
• Demonstrated experience setting technical direction across multiple teams or product areas
• Hands-on experience building and operating production AI applications, not only prototypes
• Hands-on agentic software development experience — using AI coding agents to create specs and plans, execute multi-step work across a codebase, run and fix tests, and iterate while retaining ownership of architecture and quality
• Deep expertise in either Python backend and distributed systems, or React and TypeScript product engineering, with sound judgment across the full system
• Experience with retrieval, embeddings, context, memory, evaluation, monitoring or agent architectures
• Strong grasp of security, privacy, access control, data governance and auditability for AI systems

Benefits

Gravie calls its benefits programme "the gravy". Alongside standard health and wellness cover it includes alternative medicine coverage, flexible PTO, up to 16 weeks paid parental leave, paid holidays, a 401k programme, transportation perks, education reimbursement, and two days of paid paw-ternity leave.

Location

The role is listed as Remote. Gravie's board shows most roles as Remote with some hybrid in Minneapolis, and salary bands are published in USD with a 401k and a California applicant notice — so this is realistically a United States remote role rather than an internationally open one. Confirm eligibility with Gravie if you are based elsewhere.

Why it stands out

Most listings that mention AI want you to use it incidentally. This one restructures the whole engineering role around agentic development and then raises the seniority bar accordingly. If you have been working that way already, there are not many postings that name it this directly.