Ramiro Rivas — Software Engineer
I build products end to end
Inventory systems, dealer quoting portals, and portfolio websites - with AI integrated where it actually earns its place. Eight shipped projects, each explained down to the architecture.
Selected work
8 projectsStockroom
Event-sourced inventory and invoicing engine for a window and door dealer, running on a VPS and driven by an LLM that is never allowed to do the math.
LockClaw
Compliance-first AI agent for a 550-client tax practice: WhatsApp intake, bilingual replies, and a PII firewall that no message can bypass.
Gesture Control
A Raspberry Pi camera as a touchless remote: MediaPipe hand tracking, a self-trained gesture classifier, and Wake-on-LAN over a direct Ethernet link.
Flutter Mobile Toolkit
A published Claude Code plugin: 12 slash commands and 9 specialized AI agents covering the full Flutter + Supabase mobile lifecycle.
ContractsHub
A production Flutter app for managing contracts and payments, with its own documented design system and a full testing playbook.
Dealers Portal
A .NET estimating and pricing platform for a window manufacturer dealer network: configuration-dependent pricing, discounts, terms, and quote generation.
Studio 36
A bespoke, animation-heavy portfolio built for a client: architectural designer Bryan Garcia presents his selected work under the Studio 36 name.
This Website
The portfolio you are reading, treated as a project: scroll choreography, a hand-rolled canvas simulation, and walkthrough videos recorded by a scripted browser.
About
I'm a software engineer who likes owning the whole problem: talking to the person who has it, designing the architecture, shipping the product, and running it in production. Most of the work here was built for real businesses - a window and door dealer, a tax practice, a manufacturing company - not as exercises.
I care about systems that are correct by construction: event-sourced ledgers where stock can't drift, AI agents where client PII physically cannot reach a third-party API, invariants enforced with one test each. The demos and writeups on this site go into that level of detail.
Stack-wise I move between Python, TypeScript, C#/.NET, and Flutter, and I'm comfortable from Raspberry Pi GPIO up to cloud deployment. This site itself is Astro, React, GSAP, and a hand-rolled canvas simulation - it's project number eight, writeup included.