Your systems, finally connected.

I integrate, automate and ship the software that turns isolated tools into an operation that runs on its own.

Full stackApplied AI74 certifications

  • Python
  • FastAPI
  • Node.js
  • TypeScript
  • PostgreSQL
  • Docker
  • Linux
  • Rust
  • NixOS
  • Prometheus
  • Grafana
  • OpenTelemetry
  • Sentry
  • SNMP
  • Syslog
  • VictoriaMetrics
The same systems, before and after
Disconnected tools: someone copies data between systems by hand, it takes time and some of it gets lost.Illustrative simulation. Not real data.
The problem

Every tool keeps its own version of the truth.

The CRM, the spreadsheets, the team's WhatsApp and the ERP don't talk to each other. Someone copies data by hand, and the mistakes surface once the problem has already reached the customer.

Services

Five ways to move forward.

Pick the problem that hurts today. Everything ships documented and ready for your team to run.

01

Custom development

+

The problem

An idea or a process that no off-the-shelf tool solves well.

Deliverables

  • Websites, web apps and APIs
  • Our own code, tested and documented
  • Release to production and early support
02

Systems integration

+

The problem

Each system keeps its own truth and someone copies data by hand.

Deliverables

  • Map of systems and data flows
  • APIs and connectors with retries and traceability
  • Documented data contract
03

Process automation

+

The problem

Repetitive tasks that depend on one person and fail silently.

Deliverables

  • Automated flows with failure alerts
  • Tracking dashboard
  • Operations handbook
04

Applied AI

+

The problem

The documentation exists, but nobody finds it in time.

Deliverables

  • Assistant over your documentation (RAG) with cited sources
  • Answer-quality evaluation
  • Defined limits and controls
05

DevOps and platform

+

The problem

Manual deploys, mismatched environments and surprises in production.

Deliverables

  • CI/CD pipelines
  • Infrastructure as code
  • Monitoring and alerting
Work

Technical decisions, not just deliverables.

FTTH-Copilot · NOC

From loose alerts to actionable incidents

Challenge
Many alerts per device, with no context.
Technical decision
Group by device and open an incident only when at least two distinct kinds appear.
Transferable
Cut noise in any monitoring system.
Read the case →
Python · Open source

An orchestrator with persistent context

Challenge
Chaining assisted work steps without losing the context between them.
Technical decision
An orchestrator with a CLI and persistent state, published on PyPI and covered by unit, BDD (Gherkin) and E2E tests.
Transferable
Repeatable, verifiable automation instead of loose scripts.
Read the case →
Rust · Open source

A USB driver with no official documentation

Challenge
A device with no official Linux support.
Technical decision
Reverse engineering of the USB protocol and a Rust workspace with three crates, tested in CI against several distributions.
Transferable
Integrate hardware or protocols nobody documented.
Read the case →
Vaulta · Flutter

Passwords encrypted, with no dependency on the internet

Challenge
Storing credentials without the access depending on someone else’s server, or on having a connection.
Technical decision
Derive the key from the master password with Argon2id and encrypt the vault with AES-256-GCM, which also detects any modification. Biometrics unlock it; they do not replace the encryption.
Transferable
On-device encryption when the data cannot leave the device.
Read the case →
Your case

Does your operation look like any of these?

Tell me what hurts and I’ll tell you where to start.

Write to me
Who I am

Hi, I am René Kuhm.

I have been programming since 2018 and, over those same years, I worked on an ISP network: outside plant and inside plant.

René Kuhm, founder of TecnoDespegue, at his desk

For 9 years I worked on an ISP network, on both outside and inside plant: laying and splicing fibre, drop connections, head end, OLT and ONT. First with ADSL, then throughout the migration to FTTH. Much of that work meant keeping services running that cannot be interrupted, at any hour, with a customer waiting on the other end.

In parallel, since 2018, I have devoted myself to programming. I never set one aside for the other: the same years I spent maintaining the network were the years I spent writing the software that keeps it in order. That double perspective is what I bring to every project: I know the infrastructure you touch and the code that governs it.

I work independently, so you will always deal with the same person: whoever listens to you is the one who designs the solution, writes the code and responds when something needs attention.

Every job I have taken so far came by word of mouth, with no website and no advertising. This is the first time I have written any of it down.

◇

Building software since 2018

Web apps, APIs, internal tools, automation and CI/CD.

▣

Production stack

Python, FastAPI, Node, TypeScript, PostgreSQL, Docker, Linux and Rust.

◎

9 years running networks

Outside and inside plant: fibre, head end, OLT and ONT. I know what happens when a critical system fails.

⇄

~2,500 FTTH installations

Part of the migration from ADSL to FTTH/GPON.

Training

Continuous learning, with certificates.

I study constantly. Formal training in full stack development, applied AI and networking, on top of hands-on operations experience.

74 certificates earned. These are the ones that matter most for the work.

IBM · Coursera

Full stack development

  • IBM Full Stack Software Developer (certificado profesional)
  • Node.js y Express, React, Django y SQL
  • Docker, Kubernetes y OpenShift
  • Microservicios y serverless
  • Software Developer Career Guide and Interview Preparation
IBM · Coursera

Applied AI and agents

  • IBM RAG and Agentic AI
  • Build AI Agents using MCP
  • Agentic AI con LangGraph, CrewAI, AutoGen y BeeAI
  • Agentic AI con LangChain y LangGraph
  • RAG avanzado y bases de datos vectoriales
  • Build Multimodal Generative AI Applications
Google · Google Cloud

Generative AI

  • Google AI Essentials
  • Introducción a los modelos de lenguaje grandes
  • Mecanismo de atención y modelos de transformadores (BERT)
  • IA responsable con Google Cloud
  • Escala con Google Cloud Operations
Anthropic

Claude and MCP

  • Claude Code in Action
  • Claude con la API de Anthropic
  • Claude con Amazon Bedrock y Google Vertex AI
  • Model Context Protocol: introducción y temas avanzados
  • Subagentes y agent skills
  • AI Fluency: Framework & Foundations
Packt · Google · Coursera

Networking and support

  • Networking Fundamentals & Network Access
  • Technical Support Fundamentals
  • IBM Information Technology Fundamentals
CoderHouse · freeCodeCamp · Cisco · Santander

Web and Python

  • CoderHouse: Desarrollo Web (2022), JavaScript y React (2023)
  • freeCodeCamp: Responsive Web Design y Data Analysis with Python (2024)
  • Cisco Networking Academy: Python Essentials 1 (2025)
  • Santander Open Academy: Cursor con Python, desarrollo con IA (2025)
My stack

The tools
I work with.

Languages, frameworks, data, infrastructure and observability: what I use to build, ship and run software. If your team already works with some of it, we start ahead.

Process

From diagnosis to an operation that keeps running.

01

Diagnosis

I learn your operation, map the systems and together we decide what hurts first.

02

Design

We agree on architecture, scope and success criteria before I write any code.

03

Build

Short, visible deliveries, with tests and automated deployment from day one.

04

Operations

Monitoring, documentation and handover, so the system runs without depending on me.

Contact

If this sounds familiar, let’s talk.

We copy data between systems by handWe hear about problems from the customerDeploys are scaryNobody can find the documentation

Or write to me at hola@tecnodespegue.com · I reply within 24–48 h · Eduardo Castex (La Pampa) and Córdoba, Argentina

Arrows to move, Enter to choose, Esc to close