Skip to content

About

About me

A short account of what I build, how I work, and where I am heading next.

About me

I am Satria Rangga Jati, a full stack developer working across Laravel, Vue and TypeScript on the front of a stack, and Python and FastAPI behind it. Most of what I build ends up as a working product: an interface, an API, a database and a Linux server that keeps it online.

My interest in AI is practical rather than academic. I have built retrieval systems that match a product photograph to a catalogue entry using vision embeddings and a vector index, wrapped behind a FastAPI service and consumed by a Laravel application. The interesting part is always the seam between the two runtimes.

I care about the unglamorous parts — input validation, error states, responsive layouts, accessible markup, and code that a new developer can read at 2am without guessing. Those decisions compound across every project.

Focus

What I focus on

Four categories of work, each taken from the first sketch to a running server.

Full Stack Applications

Complete products from interface design through API design, database modelling, authentication and deployment.

AI-Powered Applications

Python and FastAPI services wrapped around vision and retrieval models, integrated into Laravel backends.

Backend & API Systems

Laravel applications, REST APIs, admin panels and third-party integrations with clean, testable boundaries.

Deployment & Infrastructure

Linux server setup, Nginx configuration, process supervision and the deployment pipelines that keep it running.

Method

How I work

The same four habits on every project, regardless of stack.

  1. 01

    Understand the problem first

    Requirements and data shape before scaffolding. Most rework comes from building the wrong thing accurately.

  2. 02

    Ship the whole product

    Interface, backend, deployment and monitoring. Handing over a half-built feature is not shipping.

  3. 03

    Keep the boundaries explicit

    Clear contracts between the Laravel application, the Python AI service and the database.

  4. 04

    Optimise for the next reader

    Consistent naming, small focused classes, and code that a new developer can follow without a walkthrough.

Technology

Tools I work with

Grouped by role. No invented proficiency scores — a badge means it shipped.

Frontend

  • Vue
  • TypeScript
  • JavaScript
  • Tailwind CSS
  • Vite
  • Alpine.js

Backend

  • PHP
  • Laravel
  • REST API
  • Blade

AI / Machine Learning

  • Python
  • FastAPI
  • SigLIP2
  • DINOv2
  • FAISS

Database

  • MySQL
  • PostgreSQL

Infrastructure

  • Linux
  • Nginx
  • Git
  • GitHub Actions

Current direction

I am consolidating the AI retrieval work into something more production-shaped: faster indexing, more predictable evaluation, and failure modes that are visible rather than silent.

On the application side I am tightening my Laravel and TypeScript foundations — testing, typed contracts between the Laravel API and the Python service, and deployment that is reproducible instead of documented-from-memory.

See the work Start a conversation