A small crew
with a long memory.
RFway is a remote-native AI automation practice founded in 2016. We stay small on purpose. Every mission is run by people who've shipped production software for more than a decade — and who now spend their days deploying agents, not chatting with them.
- 2016
Founded
Two engineers, one laptop, one idea.
- 2018
First 10 missions
Established core engineering disciplines.
- 2020
Remote-native
Distributed across four continents.
- 2022
Platform era
Long-term maintenance contracts begin.
- 2024
AI core
LLM-backed agents and automations become the primary objective.
- 2026
Autonomous Observatory
AI automation for business operations is now the mission. 120+ systems in orbit.
Four rules
we actually follow.
These aren't values on a wall. They are the disagreements we keep having in meetings, which means we're still taking them seriously.
Direction Over Speed
Shipping the wrong thing fast is just expensive. We spend extra hours on scope so every hour after moves in a straight line.
Written, Not Said
Decisions, trade-offs, and postmortems live in documents you can find. If it isn’t written, it didn’t happen.
Operable By Others
Every system we build is handed to a team that isn’t us eventually. We write code as if that handoff is tomorrow.
Boring Is A Feature
We choose the unglamorous tool that’s been in production for a decade over the shiny one that launched last quarter.
The people
flying the mission.
15 years shipping systems software. Now spends his days designing agent architectures and arguing about eval harnesses.
Low-latency everything. Background in high-frequency trading systems, now building tool-use pipelines that don’t flinch.
Designs the human surface of every agent we ship. Formerly at design-led product studios in Berlin and Lisbon.
Integrations lead. Knows every way a webhook can lie to you, and wires LLMs into systems that don’t want guests.