Joaquín Minatel — Backend Engineer

About me

The short version

Backend Engineer from Buenos Aires, 21. I specialize in designing and building APIs, microservices, and production-grade data pipelines. I work with Go and Python as my main tools, and I'm comfortable with the whole surrounding ecosystem: AWS, Azure, containers, CI/CD, and AI integrations with LangChain and LangGraph.

+7

years coding

+3

years professional

Go & Py

main stack

UTN

software degree

If you want to know a bit more...

How it all started

I started at 14 with HTML, CSS, and JS. Classic. But I quickly realized that logic was my thing, not pixels. I moved to Python and never looked back.

Since then I've been self-taught: videos, official docs, courses, X (twitter), Hack The Box for cybersecurity. Always learning out of curiosity, not obligation.

Professional journey

My first job was at Dentsu as a Service Operations Analyst (May 2023, age 18). I started with L2 support on Azure and SQL Server.

In February 2024 I was promoted to Backend Engineer & DevOps. I led the automation of the invoicing flow to ERP: 20,000+ invoices a year with OCR + LLM, reducing a 30-person team's manual workload by 90%.

In parallel, I've been freelancing since 2021: APIs, AI integrations, and technical consulting for companies that need to optimize processes.

Beyond the code

I'm quite into Bitcoin and blockchain. I'm interested in finance, the stock market, and have been following BTC for years. I'm also a big F1 fan.

I'm studying a software programming degree at UTN (2025-2027). Before that I spent a year at UP studying AI Engineering, where I took math and statistics courses.

Experience

Where I've been and what I did

I design and maintain critical APIs in Go and Python (FastAPI) across Azure, AWS, and physical servers.

Finance to ERP: 20,000+ invoices a year with OCR + LLM and reconciliation against Oracle ERP. Reduced a 30-person team's manual workload by 90%.

Production CI/CD with Jenkins and GitHub Actions; automated deployments to Kubernetes via Helm and ArgoCD.

Event-driven architecture with RabbitMQ and Redis for reliable high-volume processing, with metrics and alerts.

Full-stack with Next.js (App Router, RSC, Server Actions), TypeScript, and Tailwind/shadcn when the product requires it.

L2 technical support, infrastructure administration, and databases on Azure and SQL Server.

Automated recurring tasks in Python and Bash to reduce repetitive tickets and speed up response times.

Close collaboration with development and product teams, which led to a formal transition to internal Backend Engineer.

REST APIs with Python (FastAPI, SQLAlchemy) and Go on PostgreSQL, MongoDB, MySQL, and Redis.

AI integrations: multi-step agents with LangGraph and LangChain, and document processing with OCR + LLM.

Deployments on AWS (EC2, S3, RDS, Lambda) and VPS (Hetzner, DigitalOcean), with Docker, GitHub Actions, and observability.

Maintenance of open-source templates to accelerate the bootstrap of production-grade, maintainable APIs from day one.

Projects

What I build

@joacominatel

+2,774

contributions in the last year

Repos 50+

Python

Go

TypeScript

View profile

Finance to ERP

Production at Dentsu

Critical pipeline that automates invoicing to Oracle ERP. Processes 20,000+ invoices a year with OCR + LLM, business validations, retries, and automatic reconciliation. Reduced a 30-person team's manual workload by 90%.

architecture

PDF Input OCR + LLM Extraction Validation + Score > 0.95 RabbitMQ Queue PostgreSQL / Oracle ERP Metrics + Alerts

Go Python RabbitMQ K8s Oracle

Pulse

Educational project

Real-time discovery engine. Detects emerging communities and trends by current momentum, not historical popularity. Event-driven pipeline with a 1h sliding window, low latency, and SLO/SLI defined from day one.

Event Ingestion (batch) Weight Assignment Momentum Calc (1h window) Redis Rankings REST API JWT Auth (Supabase)

Go Echo PostgreSQL Redis

FastAPI Template

Open source

Open-source boilerplate to spin up production-ready FastAPI APIs and services. Docker multi-stage, CI/CD with GitHub Actions, structured logging, and observability-oriented structure. Clean Architecture from day one.