We are a remote-first engineering team working on the unglamorous, essential problems behind production AI.
Runix is remote-first and built to work across time zones, so almost everything important happens in writing: design docs, decision logs, and postmortems rather than meetings. We are building a small, senior team where everyone owns a real slice of production and carries the consequences of their design choices. We optimize for infrastructure craft over feature churn: fewer things, kept up, measured honestly. If you like calm, deep work on systems that other teams depend on, you will probably like it here.
You run what you build. Roadmap, code, deploys, dashboards, and the pager for your area belong to you — with the authority to fix root causes instead of filing tickets about them.
We care about tail latency, failure modes, and boring, well-tested code. Slower and correct beats fast and flaky, because our customers build on top of us.
Few meetings, no performative urgency. Long stretches of uninterrupted work, async by default, and an on-call philosophy that treats a noisy pager as a bug to fix.
Metrics, incidents, and reasoning are open inside the team. We say what broke and why, internally and to customers, and we write down decisions so they can be revisited.
All roles are remote. We hire globally and default to async collaboration with a few hours of overlap.
You own the data plane of our LLM gateway: routing, provider failover, rate limiting, and the observability that makes all of it debuggable at 3 a.m. This is core-path work — every request from every customer flows through the systems you design. You will make the latency, consistency, and capacity trade-offs and live with them in production.
You might be a fit if
To apply, email [email protected] with the role name in the subject and a link to something you have built.
You own the stages of our managed data pipelines: extraction, deduplication, PII masking, and quality scoring. The job is turning messy real-world data into something a model can safely use — reproducibly, at scale, with measurable quality. You will decide where classic batch processing ends and where an LLM in the loop actually earns its cost.
You might be a fit if
To apply, email [email protected] with the role name in the subject and a link to something you have built.
You write the docs, examples, and engineering posts that help developers get real work done with our gateway and pipelines, and you talk to developers where they already are. This is a writing-heavy engineering role, not a marketing role — the output is working code and honest technical content.
You might be a fit if
To apply, email [email protected] with the role name in the subject and a link to something you have built.
Don't see your role? If you care about reliable infrastructure, write to us anyway — [email protected].