Daniel
Nogueras Del Río

CLOUD · DEVOPS · SYSTEMS

Building predictable infrastructure for uncertain systems.

Selected Systems ↓ This Site ↓ Contact ↓

Scroll ↓

42 School ◆ Cloud · DevOps · Systems ◆ AWS · Go · Linux · Terraform ◆ ES / EU

// Background

[ competition → foundations → infrastructure ]

Competition

League of Legends · Poker

Years competing at a high level taught me to make decisions under pressure, incomplete information, and constant uncertainty. That mindset still shapes how I approach complex problems today.

42 School

Systems · Low-Level Foundations

42 gave me the technical foundations to understand what happens beneath the abstractions — from memory and processes to concurrency, networking, and Unix systems.

Today

Cloud · DevOps · Infrastructure

Today I apply both perspectives to cloud and infrastructure engineering: building systems that are automated, reproducible, observable, and designed to remain predictable when things go wrong.

read the full story →

// Selected Systems

view project archive →

01 / SYSTEM

Sentinel

AI-Driven SRE Context Engine & Daemon

Go Linux Gemini API Prometheus Grafana Docker GitHub Actions

What it is

A native Go daemon designed to collect live Linux system context and transform it into structured evidence for AI-assisted SRE diagnostics.

The problem

LLMs can reason about infrastructure, but they cannot reason accurately about a system they cannot observe. Sentinel puts telemetry, processes, logs and runtime state before inference.

Architecture

01 · Host Telemetry Continuous scraping of Linux kernel metrics via Prometheus and structured system logs from journald as an immutable evidence layer.

Engineering decisions

UNIX socket communication. Keeps the CLI lightweight while the daemon maintains the live system context.

Evidence-first inference. Context is sanitized, structured and bounded before being sent to the model.

02 / SYSTEM

CloudForge

Automated Cloud Infrastructure & IaC

AWS Terraform Ansible Docker Linux GitHub Actions

What it is

A production-style AWS environment built around infrastructure-as-code, automated host configuration and reproducible containerized deployments.

The problem

Cloud environments become fragile when networking, operating-system configuration and application deployment are mixed together. CloudForge separates those responsibilities into explicit layers.

Architecture

01 · Terraform Provisioning Declarative VPC, subnets, Security Groups, EC2 instances, and least-privilege IAM roles with remote S3 state locking.

Engineering decisions

Strict layer separation. Terraform provisions infrastructure while Ansible owns host configuration and runtime deployment.

Network isolation. Internal services remain private while Nginx acts as the controlled application ingress.

02 / 02 Primary systems Proof of execution ↓

// This Site

[ infrastructure as product ]

This portfolio is not hosted on a traditional web server. It is a serverless AWS architecture designed around immutable deployments, private origins, short-lived OIDC credentials, and edge security.

GitHub Actions OIDC Short-Lived Identity

→

Terraform S3 Remote State & Locking

→

CloudFront + S3 Private Origin via OAC

→

Lambda@Edge Dynamic Security Headers

Deployment & Identity

Every deployment is triggered through GitHub Actions and authenticated against AWS using OpenID Connect (OIDC). Zero long-lived AWS keys stored in secrets.

Origin Isolation

The S3 bucket remains private and blocked to the public internet. CloudFront reaches it strictly via Origin Access Control (OAC).

Edge Hardening

Lambda@Edge applies strict Content-Security-Policy and security headers at the edge, while Route 53 and ACM ensure HTTPS delivery.

Daniel Nogueras Del Río

Sentinel

CloudForge

Daniel
Nogueras Del Río