Hugo Santos is the founder and chief executive of Namespace, a developer-infrastructure company building high-performance computing environments for software engineers and autonomous coding agents. A former Google principal engineer, he is rethinking how software gets built, tested, and integrated when machines produce code faster than conventional development workflows can absorb it.
From Portugal to Google-scale infrastructure
Santos grew up in Portugal, where an early interest in programming developed into a focus on networking and distributed systems. Around 2010, he moved to Finland to help found Blaast, an edge-computing company that delivered sophisticated applications to feature phones.
He joined Google in Switzerland in 2013 and spent nearly nine years there, becoming a principal engineer. He helped develop Google’s Boq microservices platform, a collaborative effort encompassing service development, deployment, and operations, and also worked on the Assistant platform. His account of building infrastructure at Google reflects a conviction that product teams should not have to assemble every layer of their development infrastructure themselves.
After leaving Google in 2021, Santos and his team built Foundation, an open-source application platform inspired by Boq. Its slow Kubernetes-based testing environments exposed a more immediate opportunity: faster, isolated computing infrastructure. Namespacepivoted toward specialized compute in late 2022 and acquired its first customer in June 2023. The company now provides managed GitHub Actions runners and environments for coding agents, combining virtualization, scheduling, networking, storage, and reusable build state.
Rethinking software delivery for autonomous agents
Continuous compute: Traditional pull requests and CI pipelines assume that people write changes slowly enough to tolerate delayed testing and review. Santos moves builds, tests, and validation directly inside the agent’s working loop, preserving continuous integration’s guarantees while eliminating unnecessary handoffs.
Intent-driven agent harnesses: Development begins with a concrete objective, not a pull request. An agent starts from a trusted repository state, implements the change, runs existing tests, responds to automated security or API checks, and adapts as requirements evolve. Humans evaluate whether the final outcome satisfies the original intent.
Stateful agent infrastructure: Warm caches, persistent working environments, and incremental builds prevent agents from reconstructing execution context on every iteration. Santos treats that retained state as essential to keeping validation fast enough for continuously operating systems.
Pre-merge reconciliation: Multiple agents modifying the same repository create a distributed-systems problem: proposed changes must be coordinated against a single canonical history. Santos envisions a staging layer that reconciles competing changes, groups related work, and preserves human approval before integration.
His AI Engineer Europe session develops these ideas into an architecture for agent-scale software delivery. The same engineering-first philosophy shapes Namespace’s commercial approach: Santos favors customer-facing engineers who can diagnose technical problems and deliver concrete improvements.
