Staff Frontier Engineer

Pax8

6d•$200,000 - $300,000•Remote

About The Position

The Staff Frontier Engineer role requires depth and reach together: the system is technically hard in its own right, and enough of Product & Engineering builds on it that one engineer getting it right compounds across the entire org. A representative problem is identity and entitlements, where everything depends on it and it is genuinely one of the harder problems in software. On the AI side, it is the eval infrastructure that determines whether what ships is actually prompted well. The Staff Frontier Engineer is individually accountable for building, hardening, and operating a system that the organization depends on without directly touching. They hold the problem space durably, long enough to develop real expertise and innovate within it. They earn the right to set its direction by shipping the system rather than by advising on it. This role is measured on five dimensions (Craftsmanship, Impact, Mentorship, Safety, and State of the Art). What separates it from Frontier Engineer is that Judgment, Systems Thinking, and Safety now have to hold up against a problem where being wrong is expensive and, often, no one else is positioned to catch it.

Requirements

Building: Builds and operates a system that is hard enough to need its own technical strategy and technically deep in its own right, and helps drive the technical specs for it. Builds the tool that eliminates a repeated problem.
Judgment: Makes high-stakes calls on what to build, scale, or kill when the data runs out, and escalates the ones that belong to product
Writing: Documents at a level that sustains other teams consuming the system without a standing meeting, and presents technical direction to engineering and product audiences
Outcomes: Instruments the system and proves delivered value against a falsifiable, data-backed standard
Systems Thinking: Lands fixes across ownership boundaries when the end-to-end view demands it, regardless of who owns the code
Mentorship: Grows another engineer's judgment
Safety: Applies a categorically higher bar to irreversible-consequence risk on a system with organization-wide reach
State of the Art: Owns the pipeline for adopting new techniques in the problem space
Understand AI capabilities and limitations and apply that judgment in daily engineering decisions
Use enterprise-approved AI tools (Claude) as primary productivity accelerators across the full development lifecycle
Apply Pax8 data classification and privacy principles before sharing any data with AI systems
Practice responsible AI: recognize and mitigate bias, uphold fairness and transparency requirements
Advanced prompting proficiency: multi-turn context setting, chain-of-thought elicitation, structured output enforcement, and iterative refinement
Build and own eval and validation infrastructure for AI-assisted or AI-native systems within your problem space, the layer that determines whether output is correct rather than merely plausible
Apply elevated scrutiny to agent-produced code within a highly complex system
Track state-of-the-art techniques and AI-specific vulnerability classes as they evolve, and be the one who notices when either changes and the one who ships the response into production

Nice To Haves

Track record of being the de facto owner of a problem space that most of the organization depends on, earned by building and running it rather than by advising on it
Experience mentoring other Frontier Engineers toward owning highly complex systems
Experience taking a technique from a paper or external source through prototype and proof of concept into a hardened, safety-reviewed production rollout

Responsibilities

Own, build, and operate a system that is hard enough to need its own technical strategy and technically deep in its own right, and help drive the technical specs behind it
Make consequential calls about what to build, scale, or kill on ambiguous problems where the data does not answer the question, driving them to clarity through iteration and collaboration, and own the outcome
Exercise Systems Thinking past your own ownership boundary: see and optimize the whole line end to end, and change your own plan, or push back on someone else's, when the end-to-end view says the locally optimal move is globally wrong. Pursue the real fix, by influencing, contributing, or escalating, even when it lives in a system you don't own
Own the work yourself: your domain is yours to run, so address the high-leverage problems in it, scope the solutions, and build them with a workforce of agents behind you
Instrument what you ship for adoption and usage from the start, and hold any claim of delivered value to a falsifiable, data-backed standard
Scale your own judgment into others: mentor engineers toward owning highly complex systems rather than only reviewing their output
Apply a categorically higher bar to safety, security, and irreversible-consequence risk: identify and close it out before it ships, covering both classic engineering risk (data loss, security vulnerabilities) and AI-specific risk (misuse, jailbreaks, unintended capability, data exposure), including deciding what not to ship when everything else is ready
Run the full lifecycle for adopting new techniques (discovery, spike, prototype, proof of concept, hardened production) in your problem space
Write documentation precise enough that other teams can consume your systems as a service without a standing coordination meeting

Benefits

Competitive compensation package
Annual bonus opportunity
Employee stock options
Comprehensive medical, dental and vision coverage
401(k) retirement plan with employer match
Flexible paid time off
Paid volunteer time
Wellness initiatives and employee assistance programs
Continuous learning and development opportunities
Dedicated professional development resources
Collaborative remote-first culture
Access to innovative technologies and AI-powered learning tools
401(k) with Company Match
Medical, Dental & Vision Coverage
Employee Assistance Program (EAP)
Employer-Paid Life, AD&D, Short-Term Disability & Long-Term Disability Insurance
Flexible Open PTO
Paid Sick Leave
Paid Parental Leave
RTD EcoPass (eligible Colorado-based employees)