Job description

Hi, we’re Gravie. Our mission is to create health benefits that actually benefit small and midsize businesses and their employees. Our innovative benefit solutions and services are developed and delivered by a diverse group of unique people. We encourage you to be your authentic self - we like you that way. ROLE SUMMARY Gravie is seeking a hands-on Staff Platform Engineer to do three things at once: build the golden roads and paved paths our engineering teams build on, define the north star architecture the organization moves toward, and design and deliver our AI-enabled capabilities. This is the hub role for technical direction at Gravie: long-term architectural direction and the standards that carry it converge here and radiate outward across teams. The ideal candidate combines deep distributed systems and AWS experience with real production generative AI experience, and is equally credible building a paved path teams voluntarily adopt and setting a target-state architecture the whole organization will inherit. They have established standards, reference implementations, and paved roads that engineers actually use, not just architecture documents. They will partner across Engineering, Product, Data, Security, and Infrastructure, and will work most closely with our Platform Engineering team, which is the primary vehicle for making change stick across the organization. At Gravie, engineers work closely with Product and own outcomes end to end, and we practice agentic development to increase the scope and leverage of that ownership: engineers use AI agents to help develop specs and plans, then to execute substantial multi-step engineering work, while setting guardrails and acceptance criteria and owning the quality of everything that ships. As agents take on more of the execution, the bar on problem framing, architectural judgment, system thinking, and risk identification gets higher, not lower. This is a deliberately split role: roughly 40% golden roads and paved paths (hands-on, reference implementations, shared libraries, service templates, scaffolding, pipelines - a path isn't paved until a team other than yours has shipped on it), 30% north star architecture (published decisions that get adopted and enforced by tooling, not just written down), and 30% AI solutions (agent systems, retrieval and context design, evaluation, and guardrails). The mix shifts by quarter. Being the hub does not mean being the bottleneck - a good year looks like teams making more decisions on their own because direction and the paved path are clear enough to follow without asking. The role is planned and reviewed on this split, not funded out of leftover capacity; part of the job is saying out loud when the split is breaking toward a support desk. KEY RESPONSIBILITIES - Build the paved paths teams build on, including reference implementations, shared libraries, service templates, scaffolding, and pipeline templates, so the recommended way is also the easiest way. - Treat the paved road as a product with real users: measure adoption and time to first success, and fix the parts people route around. - Do the first real implementation yourself, on a real workload with a real team, not in a demo repository. - Build the distributed systems patterns the paths encode: backend services, event-driven workers, job pipelines, APIs, and relational data models. - Partner with the Platform Engineering team as the primary delivery vehicle - build paved paths with them, plan into their roadmap, and hand off sustaining ownership of what becomes shared infrastructure. - Own migration and adoption, including moving existing services onto the path and retiring what it replaces. - Stay in code review across teams and close to production incidents as primary evidence of where the path is failing. - Define and publish the target-state architecture, with clear owners and decision dates, and keep it current as the business changes. - Own the AWS foundations product and AI workloads depend on: multi-account structure and identity boundaries, network and data isolation, compute and serverless runtimes, managed data services, and tagging and cost governance. - Drive standards adoption to completion - from documents into pipelines, with adoption measured across repositories and teams. - Lead buy, build, and reuse decisions, evaluating models, frameworks, and vendors on accuracy, latency, reliability, privacy, security, and cost evidence. - Reduce surface area as deliberately as you add it - consolidate overlapping tools and standards and own the sunset path for what gets replaced. - Design and build production-grade AI agent systems, including multi-agent workflows, orchestration layers, and supporting services. - Design retrieval, context, and memory systems that ground AI outputs, including how sensitive data is scoped, redacted, and audited. - Develop AI-powered decision-support systems meeting the accuracy, traceability, and explainability requirements of healthcare and other regulated environments. - Build product-quality user experiences using React and TypeScript where AI capability has to become understandable and actionable. - Establish patterns for AI quality assurance - automated evaluations, regression testing, groundedness checks, and compliance guardrails - so those requirements are met by default. - Establish one standard way to build, run, evaluate, and operate AI systems. - Act as the connective point across teams: surface duplicated effort, pick up decisions with no owner, and align teams heading toward the same problem. - Drive cross-team communication - writing, demos, and working sessions - so direction and tradeoffs are understood by teams not in the room. - Build alignment where incentives differ across teams, and surface disagreement rather than resolving it quietly. - Contribute to roadmap and capacity planning so paved path, foundational, and AI work is sequenced alongside product commitments. - Work through the Platform Engineering team to effect change at organization scale rather than standing up competing surfaces. - Facilitate collaboration between engineering teams, Product, Data, Security, and Infrastructure, including running forums where cross-cutting work gets planned and unblocked. - Represent technical direction to leadership and carry business context back to engineering. - Partner with Product to shape the roadmap and translate ambiguous business problems into architecture. - Raise the engineering bar through design review, mentorship, and a working decision-record practice, and make architectural context available to both engineers and agents. - Partner with Security, Compliance, and Data on AI governance, access control, and data handling, designed in rather than retrofitted. QUALIFICATIONS - Eight or more years of software engineering experience, including time with architecture scope and an organization-wide remit, and ownership of complex systems from design through production. - Still shipping: writes and reviews code today, can open a meaningful merge request in an unfamiliar codebase within the first weeks. - Experience building internal platform capability that other teams voluntarily adopted - can name the paved path or golden road built, how many teams moved onto it, and what they routed around. - Hands-on experience building and operating production AI applications, not only prototypes, including evaluation, monitoring, failure handling, guardrails, latency, and cost control. - Deep AWS expertise across a multi-account organization: identity and permission boundaries, networking, compute and serverless runtimes, managed data services, managed model services, and cost and tagging governance. Infrastructure as code in Terraform, CDK, or both. - A track record of establishing architectural foundations where little existed - specific standards introduced, how they were adopted by teams that did not report to the candidate, and which failed and why. - Hands-on experience with agentic software development - using AI coding agents to create specs and plans, execute meaningful multi-step work, run and fix tests, validate results, and iterate while owning architecture, quality, and what ships. - Deep expertise in at least one primary area (Python or JVM backend and distributed systems, or React and TypeScript product engineering), plus ability to work across the full system. - Experience with retrieval, embeddings, context, memory, evaluation, monitoring, or AI agent architectures and workflows. - Strong judgment in testing, monitoring, failure handling, privacy, security, access control, and data governance, carried into a regulated or audited environment. - Experience working with or inside a platform or infrastructure team as the delivery vehicle for organization-wide change, including how ownership was divided and how a parallel platform was avoided. - Experience serving as a central technical point across multiple teams, with evidence of making those teams faster rather than becoming their approval gate. - Demonstrated ability to create alignment across disagreeing teams, and to contribute to roadmap and capacity planning so foundational work is actually scheduled. - Clear written communication and effective collaboration across technical and non-technical teams; influence without authority is the primary tool of this role. PREFERRED QUALIFICATIONS - Healthcare, health insurance, or benefits administration, including ICHRA, claims, eligibility and enrollment, or interoperability standards such as EDI and FHIR. - Previous experience in a venture-backed high growth company. - Internal developer platform or developer experience work, including golden path or paved road programs and adoption measurement. - Experience with GitLab CI at scale, including shared and templated pipelines. - Production observability and LLM observability practice, including Datadog. - Event-driven and streaming architecture, such as Kafka or Flink. - Practical experience with agent tooling and standards such as Claude Code, the Model Context Protocol, and agent SDKs. - Experience defining an AI governance or technology radar process alongside a security or compliance partner. A LITTLE MORE ABOUT US: - We know healthcare. Our company was founded and is still led by industry veterans who have started and grown several market-leading companies in the space. - We have raised money from top tier investors who share the same long-term vision as we do of building an industry defining company that will endure over the long run. We are well capitalized. - Our clients love us. Customer satisfaction rates among employees using Gravie health plans consistently rank above 80% – nearly 40 points above the industry average. - Our culture is unique. We tend to be non-hierarchical, merit-driven, opinionated but kind people who thrive working in a high-performance, fast-paced environment. People at Gravie care deeply about making a positive impact in the lives of the people we serve. BENEFITS Our unique benefits program is the gravy, i.e., the special sauce that sets our compensation package apart. In addition to standard health and wellness benefits, Gravie’s package includes alternative medicine coverage, flexible PTO, up to 16 weeks paid parental leave, paid holidays, a 401k program, transportation perks, education reimbursement, and 2 days of paid paw-ternity leave. JOB APPLICANTS If you apply for employment with Gravie, personal information collected via our applicant tracking vendor is subject to our standalone California Job Applicant Notice at Collection, accessible directly within the application workflow and separate from this Policy.

Requirements

Must-haves

Build the paved paths teams build on, including reference implementations, shared libraries, service templates, scaffolding, and pipeline templates, so the recommended way is also the easiest way.
Treat the paved road as a product with real users: measure adoption and time to first success, and fix the parts people route around.
Do the first real implementation yourself, on a real workload with a real team, not in a demo repository.
Build the distributed systems patterns the paths encode: backend services, event-driven workers, job pipelines, APIs, and relational data models.
Partner with the Platform Engineering team as the primary delivery vehicle - build paved paths with them, plan into their roadmap, and hand off sustaining ownership of what becomes shared infrastructure.
Own migration and adoption, including moving existing services onto the path and retiring what it replaces.
Stay in code review across teams and close to production incidents as primary evidence of where the path is failing.
Define and publish the target-state architecture, with clear owners and decision dates, and keep it current as the business changes.
Drive standards adoption to completion - from documents into pipelines, with adoption measured across repositories and teams.
Lead buy, build, and reuse decisions, evaluating models, frameworks, and vendors on accuracy, latency, reliability, privacy, security, and cost evidence.
Reduce surface area as deliberately as you add it - consolidate overlapping tools and standards and own the sunset path for what gets replaced.
Design and build production-grade AI agent systems, including multi-agent workflows, orchestration layers, and supporting services.
Design retrieval, context, and memory systems that ground AI outputs, including how sensitive data is scoped, redacted, and audited.
Develop AI-powered decision-support systems meeting the accuracy, traceability, and explainability requirements of healthcare and other regulated environments.
Build product-quality user experiences using React and TypeScript where AI capability has to become understandable and actionable.
Establish patterns for AI quality assurance - automated evaluations, regression testing, groundedness checks, and compliance guardrails - so those requirements are met by default.
Establish one standard way to build, run, evaluate, and operate AI systems.

Nice-to-haves

Previous experience in a venture-backed high growth company.
Internal developer platform or developer experience work, including golden path or paved road programs and adoption measurement.
Experience with GitLab CI at scale, including shared and templated pipelines.
Production observability and LLM observability practice, including Datadog.
Event-driven and streaming architecture, such as Kafka or Flink.
Practical experience with agent tooling and standards such as Claude Code, the Model Context Protocol, and agent SDKs.
Experience defining an AI governance or technology radar process alongside a security or compliance partner.

Application prep

Use the role’s language and application details while tailoring your materials.

Keywords to mirror

aiawsreacttypescriptpythonllm

Platform: ashby
Effort: quick

Benefits

Benefits detected in the listing.

Health Insurance

Vision

401k

Pto

Parental Leave

Wellness