Lead AI Systems Engineer

Benepass

1d•$190,000 - $220,000•Hybrid

About The Position

Benepass is seeking a Lead AI Systems Engineer (P5) to build its internal AI platform from the ground up. This deeply technical and strategic role involves owning the AI systems roadmap and architecting a shared internal AI platform that starts with the engineering SDLC and scales into automation and agentic systems across the company. The role is a Staff IC position within Platform Engineering, focusing on building the platform while embedding with teams to drive adoption. Initial focus areas include engineering SDLC enhancements like coding assistants, CI/CD integration, quality automation, and knowledge systems. The platform primitives developed, such as agents, tools, retrieval, evaluations, permissions, and workflow runners, are designed to be generalizable for company-wide automation beyond engineering. Near-term success metrics include faster PR and cycle times, increased AI-tool adoption, improved test reliability, better knowledge findability, and an enhanced developer experience. Long-term success involves the platform powering durable automation outside the SDLC, such as SOPs, operational workflows, and internal systems, with demonstrable ROI and adoption. This is not a research, customer-facing product-ML, prompt-only chatbot, or manual process ownership role. It requires a Staff-level technical leader with strong platform, developer-tools, and infrastructure instincts to build full-stack AI systems end-to-end, creating AI-powered guardrails and accelerators to increase speed, confidence, and reduce toil across the company.

Requirements

Staff-level (or equivalent platform/cross-functional) impact shipping systems others depend on.
Strong programming fundamentals (Python and/or TypeScript/JavaScript preferred).
Experience building internal platforms, developer tools, or automation integrated into CI/CD.
Hands-on with modern AI systems: coding agents, LLM APIs, orchestration, and production operability.
Experience enabling quality through automation (E2E/integration testing, intelligent quality gates).
Solid systems instincts: APIs, permissions, observability, reliability.
Clear communicator who can set standards and drive multi-team adoption.

Nice To Haves

Bonus: agentic workflows, eval harnesses, RAG/knowledge systems, or ops/workflow automation beyond eng.
Bonus: fintech/regulated domains, or 0→1 platform ownership at a growth-stage company.

Responsibilities

Build Benepass’s Shared AI Platform (0→1 and 1→2)
Own the design and implementation of Benepass’s internal AI platform and strategy—engineered for company-wide leverage, not an Engineering-only toolchain.
Use the engineering SDLC as the first beachhead, while designing platform primitives that extend cleanly to Operations and other internal functions.
Stand up a pragmatic platform that combines industry-leading tools (e.g., Cursor and peers) with in-house systems, integrations, and shared infrastructure.
Define the architecture for how models, tools, context, evaluations, secrets, and permissions are composed safely inside Benepass.
Build reusable primitives: agents, tool interfaces, retrieval/context layers, workflow runners, observability, and feedback loops.
Establish the foundation for reliable environments, secure access to internal systems and data, and patterns that new domains can adopt without a ground-up rebuild.
Accelerate the Engineering SDLC with AI
Drive AI-assisted coding, review, and delivery workflows that compress time from idea → PR → production.
Integrate AI into CI/CD so quality signals, summaries, risk checks, and developer feedback show up where engineers already work.
Identify SDLC bottlenecks (local dev, code review, test wait time, release friction, knowledge gaps) and remove them with automation.
Measure what matters: PR/cycle time, adoption, developer satisfaction/DX, test speed & signal, and time-to-find knowledge.
Turn successful team-level experiments into platform defaults that scale across Engineering—and inform patterns for non-eng domains.
Enable AI-Driven Quality, Testing, and Release Confidence
Own the strategy for AI-driven test generation, maintenance, and automation—especially where it unlocks broad end-to-end coverage.
Build systems that help engineers own quality: high-signal E2E coverage, faster feedback, lower flakiness, and less manual validation.
Partner with Platform and product teams to put intelligent quality gates into CI/CD and deployment workflows.
Use AI to improve regression detection, failure triage, and the loop from requirements → test plan → execution → root-cause analysis.
Create clarity on ownership: what quality belongs to every engineer vs. what the AI/platform layer provides as shared leverage.
Build Agentic Workflows and Knowledge Systems (Eng → Company-Wide)
Design and ship agentic internal workflows that automate multi-step work, not just single-prompt assistants; starting in Engineering and expanding to other teams.
Build knowledge/search systems so people and agents can find specs, decisions, runbooks, SOPs, and operational context quickly.
Connect agents to the systems teams already use (repos, CI, docs, issue trackers, ops tools, internal dashboards) with clear permissions and auditability.
Prioritize workflows with obvious ROI: repetitive operational toil, cross-repo changes, test authoring, incident/context gathering, onboarding, and cross-functional SOPs.
Ensure these systems are observable, evaluable, and maintainable—and that the same platform can host automation outside the eng SDLC without forking the architecture.
Partner Across the Company as a Platform Force Multiplier
Operate as a Staff IC on Platform: set direction, build the core, and embed selectively where adoption and design feedback matter most.
Collaborate with engineers early so systems are testable, scriptable, and easy to integrate into existing workflows; then apply the same enablement model with non-eng partners.
Work with Engineering, Product, Design, Operations, and other leaders to choose the journeys and workflows worth automating first.
Provide documentation, reference implementations, guardrails, and golden paths so teams can adopt without heroics.
Raise the organizational bar for what “good” looks like in AI-assisted work. Software delivery first, then broader internal automation.
Define Standards, Evaluations, and Responsible Adoption
Introduce standards for AI tool usage, prompt/tool patterns, evaluation, data handling, and human-in-the-loop controls that work across Engineering and other internal domains.
Build evaluation harnesses and quality metrics so we know when AI systems are helping—and when they are creating noise.
Make pragmatic build-vs-buy decisions, favoring speed and leverage while investing in shared platform where it compounds company-wide.
Stay current on emerging coding agents, workflow agents, eval methods, and enterprise AI tooling, and bring the best of the ecosystem into Benepass deliberately.
Help Benepass adopt AI in a way that reduces toil, increases speed and coverage, and keeps humans firmly in control of quality and production outcomes.