Martino Fornasa
IT Email me

Fractional Infrastructure Lead for Startups

Build the infrastructure your team can actually run.

I help startups turn fragile, ad-hoc cloud and Kubernetes setups into declarative, observable platforms that people and AI agents can change safely. You get senior infrastructure judgment and hands-on implementation without hiring a full-time Head of Infrastructure too early.

Email me About me

Recent work

  • Hybrid ingestion platform for post-seed IoT startup. Built cloud + bare-metal infrastructure that automated onboarding for 10+ enterprise sensor providers, controlled operating cost, and was handed off to the internal team.
  • Multi-cloud / edge architecture for deep-tech startup. Designed AWS + GCP + Proxmox edge infrastructure for ingestion and inference workloads, balancing GPU access, cost, and integration with existing systems.
  • OpenTelemetry observability platform for enterprise. Replaced a licensed vendor setup with an open-source observability platform across datacenters and cloud providers, reducing license cost to zero and standardizing logs and metrics.

Frequently asked

What does 'fractional' mean exactly?

I work in parallel for a small number of clients, with continuous responsibility for a specific slice of their technical work. It is neither a one-off consulting session nor project freelancing: it's a responsibility taken over time, on a clearly bounded slice.

How many clients do you work with in parallel?

A small number, usually two or three.

How long does a typical engagement last?

It depends on the shape. A project can run a few weeks or months. A fractional engagement typically starts at three months and runs for as long as it makes sense. The goal is to no longer be needed.

Do you also implement?

Yes, I implement when it makes sense. I use agentic code tools to make that practical at this scale. Modern infrastructure (declarative, IaC-first, well-documented) is exactly the kind of codebase where these tools work. The decisions still come from engineering judgment; the implementation time goes down.

What if we need a full-time Head of Platform?

I help you define the role, evaluate candidates, and run the transition. That's part of the work, not a side effect.

Services

Agent-ready platform foundation

Early infrastructure often gets built ad-hoc. As the product grows, it needs to become robust, scalable, and safe for anyone on the team to change. Until that work is done, the codebase isn't in a state where the team, or agentic tools, can change it with confidence and speed.

AI makes changing code and infrastructure faster. But if the platform is fragile, poorly documented, or full of exceptions, it accelerates risk too. I help the team build a declarative, observable, and understandable platform so that people and AI tools can act on it with confidence.

Terraform/GitOps structure, deployment path, observability baseline, ownership model, runbooks.

Email me about this →

Cloud migration

Migrations fail on time, not on technology. Plans tend to concentrate risk at cutover, and the team is left interpreting a system they didn't build.

Migrations planned and run as IaC, so every step is automated and reversible. Risk drops by going incremental, usually starting with the biggest cost offender, which funds the rest of the work. AI-assisted analysis helps surface IaC drift, hidden dependencies, and risks before the migration begins. Provider-to-provider, hybrid, or from cloud back to bare metal: the method is the same.

AWS, GCP, Azure, bare metal, Kubernetes, AIOps, IaC, migration patterns, monitoring.

Email me about this →

Mentoring

Infra knowledge living in one head is a single point of failure. Hiring full-time leadership solves it eventually, but not soon, and not without first growing the team you already have.

I sit alongside your team for a few months at a time. Pair on real problems, review architectures, set the practices you'd otherwise learn through a year of hires and incidents. When I leave, the practice stays.

Whatever your team uses. The point isn't the tooling. It's how decisions get made.

Email me about this →

Hands-on training

When a team adopts a core technology fast (Kubernetes, containers, Terraform), parts of it end up operating without shared foundations.

Introductory, hands-on courses that bring the whole team to the same baseline on a core technology. Built around your stack, not generic exercises, so the lesson sticks. If your team is already past the basics, mentoring is usually the better fit.
See my course offerings in Kubernetes, Terraform, and cybersecurity.

Kubernetes, containers, Git, Terraform.

Email me about this →

Recent writing

  • 30 Apr 2026

    I tried Claude Design

    I tried Claude Design. For someone like me, totally hopeless at design and graphics, it's pretty impressive. In a couple of hours I sketched out a sensible visual direction for Eventitech. Then I…

  • 05 Dec 2025

    Migrating from Ingress NGINX to Envoy Gateway: Handling cert-manager integration

    The recent retirement of Ingress NGINX led a lot of people (including myself) to consider alternatives. Since we need to migrate anyway, it makes sense to migrate to the newer Gateway API , the…

All posts →

Let's talk about it

A short email with two lines of context is enough to start.

Email me