Agung Prakasya

Building tiny engines, exploring big ideas — one line of code at a time.

Background

I'm a math graduate from Institut Teknologi Bandung, where my journey into structured thinking began. With a foundation in mathematics, I naturally gravitated toward programming as a way to model complexity, solve puzzles, and build systems that behave—and scale—with intention.

What I'm Into

A lifelong coder with a soft spot for performance, precision, and elegant abstractions. I love diving into low-level details while keeping an eye on the big picture. My curiosity drives me to explore:

High-Performance Computing

Parallel processing, accelerators, memory efficiency (C++, OpenMP, SYCL)

Formal Modeling

Petri nets as a framework for AI workflows and system design

LLMs & Agents

Building tools that go beyond chat: data storytelling, visual responses, and structured reasoning

Emerging Languages

Currently tinkering with Rust, Elixir, and Vlang to stretch my mental models

Personal Projects

KnitKnot

A lightweight, embeddable property graph engine with a fluent query DSL.

Go Graph DSL design

LabKit

Lightweight homelab management for Incus containers.

Python Incus CLI tools

excvtor

An EXtremely simple CV generaTOR using LaTeX (via Tectonic). Minimalist by design.

LaTeX Go CLI tools
Under Construction

Data Storyteller

In development. An agentic LLM app that generates rich responses: tables, charts, diagrams + narrative text.

Python LangGraph Llama.cpp PostgreSQL + PgVector
Coming Soon
Under Construction

modlr

Graph & Typed PetriNet-based modeling framework. A spiritual successor to KnitKnot with deeper formal grounding.

Go formal methods Petrinet
Coming Soon
Under Construction

Homescale HPC Labs

My playground for parallel text processing using C++ and hardware like Xeon Phi, Tesla GPUs, and InfiniBand.

C++ OpenMP OpenACC SYCL HPC
Coming Soon

Learning & Exploration

I believe the best way to learn is by doing—and sharing. Recently, I've been leveling up through hands-on courses:

2024 Summer Hands-on HPC Crash Course

Oak Ridge National Lab (ORNL)

OpenMP Training Series

ORNL

Julia for HPC

ORNL

LAFF-On: Programming for High Performance

edX

High Performance Computing

Udacity

The Node.js Master Class

Pirple

Also learning: distributed systems design, agent architectures, and how to make machines "think" more clearly.

Get in Touch

If you'd like to talk about open-source, formal modeling, HPC experiments, or just geek out about Rust vs. Go—I'd love to hear from you.

No spam, no recruiters without context—let's keep it human.