Cloud Infrastructure Architect (Part-Time, Contractor)

1. Overview

La Voie's client is a consulting company headquartered in Spain, specializing in agentic AI systems and process automation for real businesses — including banking, pharma, and mid-market companies. Their mission is to analyze each client's operations and make them agentic and AI-friendly, designing and building agents, automations, and AI integrations on top of existing processes and systems, while proactively identifying opportunities for new AI products and services.

2. Job Description

As the most senior technical decision-maker after the founder, you'll design and run the cloud foundation that the company's products and client automations live on — owning the difference between a demo and a real product: control of storage, processing power, and cost.

• Evaluate AWS vs. GCP vs. cost-efficient alternatives (e.g. Hetzner VPS/bare metal) per use case, and implement a justified recommendation — hybrid setups are welcome where compliance demands a hyperscaler and cheap raw compute elsewhere doesn't.
• Design and build the core platform: environments, networking, IAM, CI/CD (GitLab), IaC (Terraform preferred), observability, and alerting.
• Own scalability and cost: storage strategy (hot/warm/cold), processing capacity for AI/agent workloads, and budget forecasting per project.
• Set up secure, isolated deployments per client, with auditability in mind given regulated-sector clients.
• Support the AI Engineer with serving infrastructure for LLM-based agents (GPU/CPU sizing, queues, caching).
• Document every architecture decision in the shared knowledge library.

3. Job Requirement

What we expect:

• 4–5+ years in infrastructure/platform/DevOps roles, with involvement in large or high-scale projects. Exceptional candidates with 2–3 years plus strong AI-era tooling mastery will also be considered.
• Expert-level cloud (AWS and/or GCP) — the founder has deep cloud expertise and will know if you're guessing.
• Terraform (or equivalent IaC), Kubernetes or serverless orchestration, CI/CD pipelines, Linux, and networking fundamentals.
• Python for automation/scripting.
• Daily, hands-on use of LLM development tooling as part of your normal workflow — a hard requirement.
• English strong enough for written specs and technical conversations.
• High autonomy — able to take a spec and deliver without hand-holding.
• Comfortable over-communicating blockers or disagreements early rather than staying silent.
• Flexibility of mind — requirements will shift as clients react, and you adapt the solution rather than defend the first idea.

What would be helpful:

• Experience running LLM/ML workloads in production.
• Self-hosted infrastructure on VPS/bare-metal providers (Hetzner, OVH, etc.) with Docker/Kubernetes, self-managed databases, backups, and monitoring without managed services.
• FinOps experience.
• Security/compliance exposure (banking, healthcare).

4. Work Conditions

• Engagement type: independent contractor (you handle your own taxes). Fully remote.
• Commitment: part-time, 15h/week (60h/month) — compatible with a full-time job. Some evening overlap with Spain (CET) expected.
• Duration: 3–6 months initial, extensible based on product and client traction.
• Compensation: €12/hour depending on seniority (€720/month).
• Start date: October 1–15, 2026.
• Tools: Jira, GitLab, Slack. No direct client contact — the founder handles all client communication.
• Documentation-first culture: all work is logged in a shared knowledge library from day one.
• Candidates who've worked together before (same company/projects) are strongly preferred.
• Interview process: a conversation with a senior expert in your discipline (no live coding) — expect to walk through real architectures you've built.