What You'll Work On

• Build and operate the runtime enforcement pipeline. It must evaluate activity from customer agents and external malicious agents, then return allow, change or block decisions with low latency and predictable failure behaviour.
• Own observability and agent tracing across the platform. Trace individual events and connect activity across internal agents, external agents, services, tools and targets. Provide useful traces, logs, metrics and alerts.
• Build the tamper-evident evidence store and audit ledger. Customers must be able to verify that decision records have not been changed.
• Enforce multi-tenant isolation in every service and data path. Use Postgres Row-Level Security, separate access paths and automated tests to prevent cross-tenant access.
• Build the infrastructure for red-team and evaluation jobs. Manage subprocesses, timeouts, cancellation and output safely. A failed job must not hang or escape its limits.
• Build reliable event and streaming services. Handle backpressure, disconnections, retries and load without losing the meaning of an enforcement decision.
• Build and operate model-serving services. Manage detector artefacts, container builds, provider routing, credentials, fallbacks and GPU workloads.
• Own cloud infrastructure and continuous delivery. This includes Cloud Run, Cloud SQL, IAM, Secret Manager and the checks that gate each deployment.
• Threat-model the infrastructure. Protect secrets, harden containers, verify artefacts and define clear trust boundaries between customer traffic, the control plane and model services.
• Oversee Product Management of software dashboard product, including UX design and design decisions.

What We Are Looking For

We need a systems engineer who takes end-to-end ownership of security-critical services. You can work from service design and data storage to deployment and production operation. You reason about latency, failure, isolation, integrity and cost. You use measurements instead of assumptions. You add tracing, metrics and logs before you call a service production-ready.

You deliver small, correct changes and test failure cases. You understand that a slow decision, a missing trace, a leaked row or a changed audit record can be a security fault. Most importantly, you want to build the systems that let organisations deploy their own autonomous systems safely and defend against malicious agents operated by others.

Qualifications

We assess demonstrated ability, rather than a specific credential or number of years in industry. You should have:

• Strong experience with backend, distributed or infrastructure systems. Experience with Python is preferred.
• Production experience with observability and distributed tracing. You can connect actions, tool calls, requests, model calls and enforcement decisions across agents and services.
• Experience building and testing responsive services under load. You understand latency, timeouts, retries, backpressure and failure handling.
• Experience with relational databases and multi-tenant data. This should include Postgres, Row-Level Security, migrations and access-control tests.
• Hands-on experience with containers, cloud infrastructure and continuous delivery. Experience with GCP, Cloud Run and Cloud SQL is preferred.
• A good understanding of operational security. This includes least privilege, secret management, hardened containers and trust boundaries.
• Experience with UX design and Product Management is preferred.

Particularly strong signals include:

• A production system that you owned where latency, correctness or integrity under load was important.
• Production observability or agent-tracing work, including the correlation of multi-step or multi-agent activity with OpenTelemetry or similar tools.
• Experience with multi-tenant SaaS isolation, Postgres Row-Level Security or compliance-driven infrastructure.
• Model-serving or MLOps experience, including GPU inference, detector artefacts, provider routing or containerised model deployment.
• Work with sandboxing, subprocess isolation, tamper-evident logging, supply-chain integrity or other security controls.
• Open-source infrastructure contributions or systems write-ups that show clear engineering decisions.

We do not expect any candidate to have worked across all of these areas.

What We Offer

• Competitive salary, meaningful founding equity, and a performance-linked bonus.
• Direct influence over the product, its architecture, and company direction.
• Access to real, high-consequence AI deployments rather than purely synthetic environments.
• The opportunity to establish Refractal's systems engineering function and grow into a senior technical leadership role.
• The chance to work directly with founders whose experience spans MIT, NASA, Microsoft, government, venture capital, and early-stage technology companies.
• London-based, with an in-person culture. We sponsor UK Skilled Worker and can assist with Global Talent visas.

How to Apply

Email careers@refractal-ai.com with your CV and a short note on the most interesting system you have built, scaled or secured. Links to open-source work, technical write-ups or systems you have designed are strongly encouraged.

Our process is fast: an intro call with the founders, a technical deep-dive on your past work, a short practical exercise, and an offer, typically within two weeks.