Principal Systems Architect

Sovereign AI & Infrastructure

Location: Remote — anywhere with reliable internet Overlap: 4+ hours daily with Central European Time (CET/CEST) Travel: Biannual team summits in Europe, fully funded Reports to: Founder / CTO

The Opportunity

Most infrastructure engineers spend their careers configuring someone else's computer. This role is different. You will design and build the computer.

We're hiring a Principal Systems Architect to lead the core infrastructure behind Nuvolos, a platform that gives universities, central banks, and research institutions full sovereignty over their computing environments. Where public cloud providers ask institutions to hand over their most sensitive data and hope for the best, we give them a sovereign workspace they actually own — one that can capture the complete state of any scientific environment, move workloads seamlessly between dedicated hardware and cloud burst capacity, and run autonomous AI agents entirely within a secure perimeter.

This is a foundational role. You'll report directly to the founder, shape the technical direction of the platform, and work on problems that sit at the intersection of bare-metal infrastructure, Kubernetes orchestration and distributed storage — all in service of scientific discovery and institutional independence.

About Nuvolos

Nuvolos (nuvolos.com) is the Unified Workspace for Sovereign AI & Science, built by ALPHACRUNCHER in Europe. Our clients include central banks, life-sciences research groups, and economics departments at leading universities. We address what we call the "Blind Confidence in Cloud" crisis: the assumption that handing institutional intelligence to hyperscalers is safe, affordable, and reproducible. It isn't.

Our platform captures the full state of any computing environment — operating system, code, data, and results — so that a drug discovery simulation run today can be reproduced bit-for-bit a decade from now. Workloads run on dedicated bare-metal hardware for performance and compliance, but can burst seamlessly to public cloud when massive scale is needed. We integrate AI-native coding environments and local LLM hosting so institutions can deploy autonomous AI agents without their proprietary data ever leaving the sovereign perimeter.

We're a small, focused team. You will not be a cog in a machine. You will be the architect of the machine.

What You'll Do

Your mandate is to eliminate the friction between sovereign control and agentic agility:

Architect the sovereign infrastructure layer. Design and operate Kubernetes across bare-metal, on-premises, and managed-cloud environments. Our stack is built on Talos Linux with a Cozystack-based orchestration layer. You'll own the provisioning, lifecycle, and networking of these clusters — and push the boundaries of bare-metal Kubernetes together with our enterprise support partners.

Own the state-capture engine. Our platform captures the full recipe of any computing environment: OS, code, data, results, and — increasingly — agent state (model versions, system prompts, reasoning traces). You'll extend this system to ensure total reproducibility of AI-assisted research over decade-long timescales.

Design the hybrid bursting architecture. When a researcher needs 200 GPUs for 48 hours, the platform should scale seamlessly from dedicated hardware to public cloud and back — without the user noticing, and without the finance office panicking. You'll build the orchestration and budget-enforcement logic that makes this possible.

Implement resource governance. Institutional clients operate with complex funding structures: a single research project might draw compute from one grant and storage from another. You'll translate these financial realities into infrastructure-level resource management and billing logic.

Ship infrastructure as product. You'll collaborate directly with the founder to turn infrastructure capabilities into features that close enterprise deals.

You Might Thrive Here If…

We care about what you can do and how you think, not how many years appear on your CV. That said, this is a principal-level role and requires deep, hard-won expertise.

You've built and operated Kubernetes in multiple environments — managed clusters, on-prem deployments, and ideally bare-metal setups.

You understand infrastructure at the hardware level. You've worked with switches/routers, configured virtual or physical networks or configured a firewall. You are interested in how Cloud services work and how they can be operated on Kubernetes.

You've worked with/operated distributed storage systems. Ceph, CephFS, GPFS, Lustre, or similar. You understand replication, erasure coding, and what happens when a storage node disappears at 2am.

You’ve worked on scaleable products. You have experience in how the development, testing, building, deployment and monitoring of modern IT products is done in Kubernetes or using cloud services.

You write clearly and communicate asynchronously. In a remote-first team, your RFCs, architecture decision records, and pull request descriptions are your primary leadership tool. You write to persuade, to clarify, and to create institutional memory.

You've worked in regulated or high-trust environments. Banking, healthcare, government, defence, or academic research. You understand why "just use S3" isn't an acceptable answer when a central bank asks where their data lives.

You're product-minded. You can translate a CTO's business vision into technical architecture and a client's compliance requirement into an infrastructure feature. You think about what to build, not just how.

You'll Stand Out If…

You've worked with immutable Linux distributions (e.g. Talos) or built Kubernetes clusters from scratch without managed services.
You have experience with air-gapped or disconnected deployments where nothing can call home.
You've designed GPU orchestration for HPC or scientific computing workloads — memory bandwidth optimisation, NUMA-aware scheduling, or multi-GPU training pipelines.
You've contributed to open-source infrastructure projects (CNCF ecosystem, Ceph, Kubernetes operators, or similar).

The Stack

You won't be expected to know all of this on day one. But this is the landscape you'll operate in:

Orchestration: Kubernetes across managed, on-prem, and bare-metal environments; Talos Linux; Cozystack for sovereign deployments Storage: Ceph/CephFS, Snowflake, PostgreSQL, Redis Languages: Python (backend services), JavaScript (frontend), Shell scripting (infrastructure automation) AI Tooling: GitHub Copilot, Claude Code, Google Antigravity Observability: Prometheus, Grafana CI/CD: Helm charts, GitHub Actions

How We Work

We are a remote-first team. This is not remote-friendly or remote-tolerated — our processes, decisions, and communication are designed for distributed work from the ground up.

Async by default. Most collaboration happens in writing: architecture docs, PR reviews, message boards, and structured decision records. We optimize for thoughtful async communication over synchronous meetings.

Overlap with CET. Our core team operates in the Central European timezone. We ask for a minimum of 4 hours of daily overlap with CET (roughly 9:00–18:00 CET) to enable real-time pairing, incident response, and the kind of rapid back-and-forth that complex infrastructure work sometimes demands.

Biannual team summits. We gather in person twice a year, typically in Budapest or elsewhere in Europe, for 4–5 days of strategic planning, architecture deep-dives, and the kind of creative work that benefits from being in the same room. Travel and accommodation are fully funded.

Tools. GitHub for code and project management, Slack for quick coordination, long-form docs for everything that matters. We don't do daily standups or status theatre.

Compensation & Benefits

We believe in transparency. Here's what we offer:

Competitive base salary: Depending on experience and what you bring to the table. Our compensation is benchmarked against European infrastructure startups.

Meaningful equity: Employee Stock Ownership Plan (ESOP) with a meaningful stake in the company. You're joining early enough that your architecture decisions will directly shape the platform's value. We want your incentives aligned accordingly. Details shared in initial conversations.

Hardware budget: We'll equip you with whatever you need to do your best work.

Learning & conferences: Annual budget for professional development, conferences (KubeCon, FOSDEM, relevant academic conferences), and certifications.

Flexible time off: We trust you to manage your own schedule and rest. No artificial caps.

How We Hire

We respect your time. Our process has four stages, and we aim to complete it within 3–4 weeks:

Application review. Send us your CV and a short note (a few paragraphs, not a cover letter) explaining what about this role interests you and what you've built that's relevant. We read every application. No AI screening.

Founder conversation (60 min). A mutual exploration with the founder/CTO. We'll discuss your background, the platform's technical challenges, and whether the mission resonates. We'll answer every question you have.

Technical deep-dive (90 min). A structured technical discussion. We'll work through a real architectural challenge from our domain — think "how would you design the burst-to-cloud orchestration layer?" — and explore your reasoning, trade-offs, and instincts. No trick questions. No binary trees.

Team conversation (45 min). Meet the people you'd work with. This is as much for you as it is for us.

We aim to give you a decision within 5 business days of your final conversation. If we're not a fit, we'll tell you why.

A Note on Applying

You may not meet every item on this page, and that's fine. If this work excites you and you bring relevant depth in some of these areas, we'd genuinely like to hear from you. The best infrastructure teams are built from people with different backgrounds, experiences, and perspectives — and we're committed to building one of those.

We welcome candidates from non-traditional paths. If you learned distributed systems by running game servers, or you understand bare-metal provisioning because you built a home lab that got out of hand, that counts. Show us what you've built.

Ready to build the infrastructure for sovereign science? Apply via email at careers@alphacruncher.com