auttapon.

Engineer · Builder · Curious mind

Auttapon.

Senior Software Engineer

Interesting problems.
Useful things.

I turn ideas into working software with AI as a development partner—and engineering judgment to guide the work.

01 / CreationsThings I build Competitive games. Playful learning. Community tools.

Selected work

MathArena

An online equation board game with real-time multiplayer, bot practice, rankings, and tournaments.

Real-time systems · Game rules · Product delivery

Behind MathArena

The challenge

Correctness goes beyond arithmetic. Players need consistent turns, scores, and clocks—even when a connection drops.

The engineering

The server validates tile placements and every equation they form. Redis holds live match state; Postgres stores match records. Room locks coordinate competing actions, while regression tests cover reconnect and spectator recovery.

The tradeoff

Separating live and durable state adds coordination and operational work. Consistency across those boundaries needs explicit handling and testing.

Pixel Arena

A playable pixel RPG that teaches equation-building through quests, board puzzles, and a progress-driven world.

React + Phaser · Rule validation · Learning UX

Behind Pixel Arena

The challenge

Make a complex equation game approachable through play, with lessons that reflect the rules players will use in competition.

The engineering

React handles lesson UI and accessible controls; Phaser handles the world and movement. Typed events connect them. Rule-parity fixtures check against the competitive engine, and content validation checks legal solutions, routes, and progression.

The tradeoff

A local rule snapshot lets the learning app deploy independently. That freedom comes with a responsibility to keep its rules and fixtures aligned with the main game.

Neighborhood Flood Watch

A mobile-first community dashboard combining water levels, road reports, public cameras, and weather—with data freshness made visible.

Public data integration · Cloudflare Workers · Community UX

Behind Flood Watch

The challenge

Bring scattered information into a useful view for residents, without hiding differences in source, location, or update time.

The engineering

The interface brings road reports, cameras, water levels, and forecasts together with visible sources and timestamps. The initial data layer distinguishes missing readings from zero, checks freshness, and uses a small Worker with fixed upstreams, caching, and timeouts.

The tradeoff

The information depends on external feeds and local reports. Freshness and uncertainty remain part of the experience; nearby readings alone cannot establish road conditions at a specific location.

02 / Under the hoodLook a little closer Simple on the surface. Intentional underneath.

Human-led, AI-assisted

I use AI to accelerate development, while keeping engineering decisions, validation, and delivery under human ownership.

Clear boundaries, rule fixtures, and automated checks make the work easier to verify and continue.

This website: plain HTML, CSS, and a little JavaScript. Made for Cloudflare Workers.

03 / ConnectionsStart a conversation Work, technology, or an idea worth trying.

Let’s exchange ideas

Have a good problem or an idea worth exploring? Find me here.