Member of Technical Staff, Infrastructure

Obvious

4d•$220,000 - $280,000•Remote

About The Position

We're building an AI-native workspace—an operating system for work that puts co-intelligence at the center. Start with data or an idea, describe your goal, and Obvious goes to work: running analysis, searching the web, writing documents, generating tables, designing presentations, visualizing data, building dashboards, and more. As Steve Jobs imagined the personal computer as a bicycle for the mind, Obvious imagines AI as a garden for the mind. Less mechanical acceleration. More organic cultivation. What if, instead of just vibe coding, you could vibe-work? What if getting from idea to done wasn't so opaque, stubborn, and high-latency? What if there was a way to consistently deliver work that feels like it came from the best version of you on your best day? That's Obvious. We're not looking for the traditional IT-professional profile—someone who knows Linux, box configuration, and enterprise DevOps but hasn't rethought infrastructure for the AI era. We're looking for an engineer who has spent their career treating infrastructure as the product itself, and who brings that lens to building AI-native infrastructure tooling. That means owning the systems that make every Obvious engineer, and every Obvious agent, more productive: build and deploy pipelines where rolling back and forth is trivial, model-serving and inference infrastructure that holds up under real AI workloads, and the observability to know what's actually happening in a system that's non-deterministic by nature. We are small and talent-dense. Among our founding team, we have world-class builders, former founders, and leaders from companies like Netflix, Google, Uber, Meta, Dropbox, Instacart, Shopify, Apple, Datadog, and Twitter (X). If you're excited to build infrastructure that enables others—human and agent—to do their best work, join us.

Requirements

Come from a company where infrastructure was the product itself—not infrastructure work done in service of someone else's product. Think platforms like Vercel, Railway, Fly.io, Render, Heroku, Netlify, Supabase, Modal, or similar—ideally as an early hire or in a role with real ownership
Applied that infra background specifically to AI/LLM workloads: model serving, inference infrastructure, agent tooling, eval pipelines, or LLM-ops guardrails—working at an "AI company" alone doesn't count if the infra work itself doesn't show this lens
Real technical depth in distributed systems: Rust or Go, storage engines, control planes, Ceph, RDMA, eBPF, bare-metal automation, or Kubernetes internals (not just usage)
A proven record of exceptional achievements and impact
Strong Terraform skills—you've managed real infrastructure as code
Hands-on experience with observability tools: OpenTelemetry, Datadog, Dash0, Braintrust, distributed tracing, metrics, structured logging
You've been on-call, and you've built systems that made on-call better
You think like a product manager for internal tools, where the product is developer (and agent) productivity
Willingness to work hard, move fast, and grow quickly in a rapidly changing environment
A humble attitude, an eagerness to help your colleagues, and a desire to do whatever it takes to make the team succeed

Nice To Haves

A personal or self-directed infra track record: side projects, homelabs, open-source infra tooling, published writing or talks—signal of genuine intrinsic interest, not just job history
Security chops: IAM, zero-trust, secrets management
SRE practices: SLOs, SLIs, error budgets, chaos engineering
Cost optimization for cloud infrastructure
Based in Atlanta (nice to have, not required)
You love the talk "Simple Made Easy"

Responsibilities

Make deployments boring (in the best way possible)
Own CI/CD pipelines: optimize build times, improve caching, reduce flakiness
Evolve our Kubernetes (EKS) deployment strategy for reliability and speed
Build and harden the infrastructure behind model serving, inference, and agent tooling—not just the app layer around them
Extend our telemetry with better instrumentation, smarter sampling, and actionable dashboards, including eval pipelines and LLM-ops guardrails
Build alerting that catches actual problems and ignores noise
Make the feedback loop from code to production as fast as possible
Improve preview environments, local dev tooling, and testing infrastructure
Eliminate toil through thoughtful automation, not another dashboard nobody reads
Be the engineer who makes other engineers—and agents—faster