Rosyidi

Software engineer with a foundation in physics and data science.

Jakarta, Indonesia

Scroll

LAYER 01Background

The Foundation

A foundation in physics.

Built numerical simulations to model how electrons behave within quantum systems. That discipline established the method that still governs the work: decomposing complex problems into parts a machine can solve.

LAYER 02Approach

The Craft

Focused on the backend.

Build and maintain the systems behind digital commerce, where a single dropped order carries real cost. Work spans the full stack, with attention on the queues, caching, pipelines, and architecture that hold steady under load.

DIM 01● Core

Software Engineering

The core. Backend systems, marketplace integrations, and architecture built to scale.

DIM 02

Data Science

Machine learning pipelines and spatial analytics that turn raw data into decisions.

DIM 03

Computational Physics

Numerical methods and simulation, the analytical bedrock beneath the work.

// technical-stack

Backend

TypeScriptNode.jsHapi.jsFastifyElysiaASP.NET CoreGolang

Frontend

React.jsNext.jsTailwind CSS

Data & ML

PythonTensorFlowScikit-learnPandasNumPyFortranMATLAB

Infrastructure

PostgreSQLMongoDBRedisClickHouseDockerKubernetesGCP

GIS & Spatial

PostGISLeaflet.jsMapbox GLQGIS

LAYER 03In practice

The Work

Where analysis becomes systems.

marketplace channels integrated

orders synced / peak event

100K+

geospatial polygons processed

0.94

R² · bandgap ML model

// experience

Jubelio

Middle Software Engineer

2023 — Now

Part of the Service team in a 4-person squad responsible for marketplace integrations across 12 channels (Shopee, TikTok Shop, Lazada, Blibli, and others). The team owns the SDK consumed by all other internal teams and acts as the primary liaison between Jubelio and marketplace partners.

Atom Daya Utama

Software Engineer

2021 — 2022

Built GIS-based products from scratch as a fullstack engineer, working in a small team with direct involvement from the CEO and manager.

INDONESIAN INSTITUTE OF SCIENCES (LIPI)

Research Assistant

2020 — 2020

Reproduced and extended the methodology from Zhuo et al. (2018) on bandgap energy prediction using a different dataset of inorganic compounds. Applied RFE to reduce 286 features down to 65 most significant material properties, then benchmarked multiple models. Random Forest achieved the best result at R² = 0.94.

// featured-projects

N01