i'm a software engineer based in the philippines, driven by curiosity and whatever i find interesting. i learn fast with the caveat of goldfish memory (which i call my context lol) and i'm usually self-motivated that i forget to eat.
this is not like other portfolios. feel free to have a look around and explore what makes me, me.
drag empty cork to move around.
drag notes to rearrange them.
use +/− or pinch-wheel to zoom.
click any note to open it.
@schmayterling
open githubrecruiter?
(follow the red line) →the same circle, over and over.
for those who want something simpler
enter simple mode
i enjoy writing code.
if ai solves every problem,
what's left?892 commits. the personal AI agent i've been trying to build since 2023 is finally becoming something i'm proud of.
one question usually turns into three projects and too many open tabs.
typescript, python, rust, go, infrastructure, and agents.
in roughly five weeks
an earlier role handled 2,000+ requests a day at 99.3% uptime.
i update this board whenever my attention changes, so it never settles into a final form.
how WFDs work, when to write one, the states a document can be in, frontmatter format, and the markdown features available.
my desktop is scattered screenshots, dozens of windows, half-finished checklists, and constant switching.
whatever feels urgent or interesting gets the whole pile.
right now, only this website and my ai work count as current.
did i mention this board evolves over time? you might find new stuff here next time. suggestions? email me.
hey@ryanaque.com
i turned weeks of neuroscience papers into her memory system.
weave gives long-running agents memory, state tracking, and retrieval that can be inspected.
i watched a kdrama about a girl who forgets everything every morning and realized my AI agent has the same condition. so i built it a brain.
someone else is
on the other end.
i'm less interested in whether it's good.
ideally when something is coming up, whether it's work related or it's just a personal project of mine i usually let it run at the end because i need pressure to perform things.
diffusion models
i built an AI agent with persistent memory. now i need to figure out if it's remembering the truth or just what i wanted to hear.
AI makes me faster. It also might be making me dumber. I'm figuring out where the line is.
“maaaaaaaaaaay 😭”
— nin0a record of the local markdown editor i built around this site's custom rendering engine. the public experiment has since been retired.
atlas maps symbols, calls, tests, and history so developers and agents can ask structural questions about a repository.
2,239-trial benchmark across 8 OSS repos: Atlas beats a text-search baseline by +0.223 deterministic, +0.127 LLM-judge, at 42% fewer tokens.
orchestrator lets a local model read and control a live chromium page through the devtools protocol.
opening note
wfd / Jul 21, 2026 / 5 min
i write WFDs in markdown. the workflow has been: edit the .md file in my editor, save it, wait for next.js to recompile, switch to the browser, scroll to wherever i was, check if it looks right, switch back, fix whatever's off, save, wait, switch, scroll, repeat.
the friction adds up when you're iterating on formatting, especially with the custom syntax this site uses. admonitions, media directives, mermaid diagrams, component embeds. you can't preview any of that without running the full build pipeline.
i looked at alternatives. HedgeDoc is great for collaborative editing but doesn't understand my custom syntax. StackEdit and Dillinger are standard markdown editors that render GFM and nothing else. none of them know what @video[caption](url) or @component[PixelLoader] means, because those are extensions i built into the rendering engine.
so i built my own.
Retired the public flow editor experiment has been retired. this post and its images remain as a record of how it worked.
a side-by-side markdown editor that runs entirely in the browser. editor on the left, live preview on the right. the preview uses the exact same rendering pipeline that produces the WFD posts you're reading right now: remark, rehype, shiki, and all the custom block parsing on top.
nothing leaves your browser. the markdown is parsed client-side, stored in localStorage, and never sent to a server. when you're done, you download the .md file to your machine.
the editor debounces input and re-renders the preview after you stop typing. the toolbar shows render time in milliseconds, whether parsing succeeded, and the current diagnostic count.
the preview doesn't render frontmatter headers yet. i'll update this once i get around to it.
typing / at the beginning of a line opens a command palette with all available syntax. /note inserts an admonition template. /mermaid inserts a diagram scaffold. /component inserts a component directive. the list filters as you type and you can navigate with arrow keys.
this was the main motivator for building the editor. i kept forgetting the exact syntax for things like admonition titles (> [!NOTE/Custom Title]) or component props (@component[Name]({"key":"val"})). now i just type / and pick from the list.
the editor runs a linter that catches structural issues in real time. unclosed code fences, malformed media directives, invalid component JSON, unknown admonition types, headings that are too deep. errors and warnings show up as colored line numbers in the gutter.
hovering a diagnostic line shows a rust-style error tooltip with the offending line, caret markers, and a help hint when available:
error[E]: unclosed code fence
--> editor:12
|
12 | ```typescript
| ^^^^^^^^^^^^^^^^
= help: this code block starting here is never closed
errors with severity error also surface as styled containers at the top of the preview pane, so you don't have to hover individual lines to spot problems. blocks that fail to render are silently dropped rather than showing inline error boxes because the top-level error banner already tells you what's wrong.
clicking an image embed in the preview opens it in a fullscreen lightbox with prev/next navigation (matching the behavior on the live site). clicking a mermaid diagram opens a draggable canvas where you can pan by dragging and zoom with the scroll wheel or the +/- keys. the canvas shows zoom controls in the top-left and an × button to close. pressing 0 resets the view.
.md file. also bound to ⌘S / Ctrl+S..md, .markdown, or .txt files.the split between editor and preview is resizable by dragging the divider. there's also a toggle to hide the preview entirely, which centers the editor at US Letter width (816px) for focused writing.
on mobile the editor and preview are separate tabs since side-by-side doesn't work on small screens.
the gutter shows line numbers synced to the editor's scroll position. diagnostic lines are highlighted in red for errors and yellow for warnings.
vscode has a built-in markdown preview, and there are extensions for GFM and custom renderers. none of them would use my exact rendering pipeline without significant effort to replicate it. the custom block parser, the admonition component, the media carousel logic, the shiki highlighting config: all of that lives in the site's codebase. the flow editor imports the same function the site uses, so the output is identical.
the rendering engine, the editor, and the WFD system are all built for this site specifically, but the patterns are generic enough that someone could adapt them. if this blog gets any traction or if there's interest, maybe i'll open source the whole thing. the markdown engine alone might be useful to someone building a similar system.
if you have thoughts on that, let me know.