Careers

Not a typical startup.
Enterprise infrastructure.

Paying customers from day 1. Real problems, not experiments. Measurable impact (not vanity metrics). Cutting-edge tech (not CRUD). Team with 45+ years combined experience from Google, AWS, Fortune 500.

We're looking for people who've built systems in production, dealt with enterprise scale, and understand that "works" ≠ "ready".

Explore Roles Get in touch
Decision Audit Log
A. Reyes · Platform Eng 09:14:22
Released context-engine v2.4.1 after 3-gate review
GATE PASSED · IMMUTABLE
D. Park · AI Safety 09:31:07
Flagged drift +0.18σ in prod agent, runbook triggered
INCIDENT OPEN · RUNBOOK-07
S. Mbeki · Infra 10:02:55
Budget guardrail triggered, $12k overage blocked
POLICY ENFORCED · LINEAGE CAPTURED
M. Chen · Research 10:18:40
Eval suite added for red-team scenario 12
REGRESSION GATE UPDATED
You · ? — : — : —
Your entry here.
AWAITING CONTRIBUTION
EVERY DECISION HAS A TRAIL · END-TO-END
High stakes
Real consequences
Governance, reliability, traceability. Not CRUD, problems that have real consequences.
Low handoff
Full-stack ownership
Signals to impact, you own the full loop.
No fluff
Craft and rigor
No sprints-for-the-sake-of-it. No LeetCode. Uncompromising quality and systems thinking only.
Work

What you'll work on

Decision Infra
Gates, policies, approvals, evidence linking Audit exports, lineage tracing, structured gate enforcement across organizations with complex approval chains.
Agent Systems
Governed tool use, full traces, rollback-ready workflows Maximum autonomy under hard policy constraints. Agents that stay in bounds with gates and auditability.
Data & Signals
Transformation pipelines, quality gates, real-time consistency Versioned lineage, source-of-truth signals feeding decisions at production scale.
Safety & Eval
Release gates, red teaming hooks, regression suites Policy enforcement at scale. Systems that prevent regressions before they reach production.
AI Operations
Drift detection, incident runbooks, budget guardrails Production monitoring and reliability for AI that runs in the real world, not just demos.
Culture

How we work

Clarity beats urgency. We optimize for correct, defensible outcomes, not speed at all costs. A decision made under pressure without structure gets made twice.
Ownership with accountability. Every decision and every line of code has a trail. You own the things you build end-to-end, from design through production.
Systems thinking. Every feature is infrastructure, not a one-off. We design for the second and third-order effects from the start.
Write things down. Specs, gates, postmortems, learnings. Knowledge compounds when it's captured. We don't rely on tribal memory.
Process

How we hire

No trivia. No LeetCode. A realistic, respectful process that gets to signal fast and doesn't waste your time.

“We hire for how you think about systems, not how you solve puzzles under a timer.”

Zirvox Hiring Team
1
Intro call
Fit, context, and mutual interest. 30 minutes, no homework required.
2
Technical interview
Role-specific deep dive on craft, systems thinking, and how you approach hard problems. 60–90 minutes.
3
Work sample
A realistic problem in your stack. Not trivia. Not LeetCode. Something you'd actually encounter here.
4
Team loop
Collaboration, ownership, and systems thinking across 2–3 sessions with the people you'd actually work with.
5
Offer + references
When we're aligned we move fast. References when needed, not by default.
Join

Build something that matters.

We're early, deliberate, and building for the long term. If that sounds like your kind of place, we'd like to hear from you.

Email us About Zirvox