Who is Lucien George?

Lucien George is a Senior Product Engineer at Fyxer, based in London and originally from Beirut, Lebanon. He builds products end-to-end and previously worked at Shopify, Le Wagon, and co-founded Localista, Skyla, and Impact Lebanon.

Is it Lucien George or Lucien Georges?

The name is Lucien George, no s. It gets written Lucien Georges or Lucian George often enough that both turn up in search, and George Lucien reverses it, but all three are wrong: it is Lucien George.

What does Lucien do at Fyxer?

Lucien is a Senior Product Engineer leading development of Fyxer's notetaker product - a native macOS and Windows desktop app (Electron) that records meetings in the background without a bot. The product reached 1,000 weekly active users within months of launch.

Where is Lucien based?

Lucien is based in London, UK. He grew up in Beirut, Lebanon, and lived in Montreal, Canada while studying software engineering at McGill University.

What is Lucien's tech stack?

Lucien's primary stack is TypeScript and React with the TanStack ecosystem (Start, Router, Query, Form), Tailwind CSS, shadcn/ui, and Convex on the backend. He has deep experience with Ruby on Rails, Python, Electron, and native mobile (Swift/Kotlin/React Native).

Where did Lucien study?

Lucien holds a Bachelor of Engineering in Software Engineering from McGill University (2013–2018), with an exchange semester at UNSW Sydney. He also attended Le Wagon London (Batch #190, 2018) and Harvard Business School's Families in Business program (2022).

What languages does Lucien speak?

Lucien speaks English, French, and Arabic fluently.

Lucien George | Developer and agent resources

This site is built to be read by software as well as by people. Everything below is a stable, public URL. Nothing here needs a key, and nothing here is rate limited beyond ordinary abuse protection.

Machine-readable surface

/llms.txt: the link index. Start here to see what exists.
/llms-full.txt: every section's markdown concatenated into one file. The cheapest way to ingest the whole site in a single fetch.
/index.md: the agent entry point for the site root, with the page map and guidance on when this site is the right source.
/agents.md: instructions for an agent, including when to use this site and when to look elsewhere.
/sitemap.xml: every canonical URL with a lastmod.
/robots.txt: the crawl policy. AI crawlers and user-triggered agents are named and allowed; only /api/ is closed.
JSON-LD (schema.org Person, Organization, WebSite, FAQPage) is inlined in the <head> of every HTML page.

Markdown for every page

Every HTML page has a markdown twin at the same path plus .md. For example /work/fyxer is also /work/fyxer.md.

The canonical URLs also do content negotiation, per acceptmarkdown.com and RFC 9110. Ask for markdown and you get markdown from the same URL you found in search:

curl -sI -H "Accept: text/markdown" https://www.luciengeorge.com/about
# content-type: text/markdown; charset=utf-8
# vary: Accept, Accept-Encoding

Accept is ranked properly (q-value, then specificity), so a browser still gets HTML. Every markdown response carries YAML frontmatter with title, description, and the canonical url, so a fetched file can be attributed back to its page.

A 404 is answered the same way. Any client that did not explicitly ask for HTML gets a markdown body naming where to look next, rather than an app shell with a 200 on it.

Source code

The site is open source: github.com/luciengeorge/lucien. It is a TanStack Start app on Vercel with Convex behind it, and the assistant answers out of a RAG index built from the same markdown that renders these pages. The repository carries AGENTS.md and CODEBASE_ARCHITECTURE.md at its root, which is what a coding agent should read before changing anything in it.

The article A portfolio that answers questions about me, gated by an LLM judge explains how the retrieval and the eval gate work, with the parts that turned out to be wrong left in.

What does not exist

There is no public API, no OpenAPI specification, no MCP server, no SDK, and no webhook. This is one person's portfolio, so there is nothing to integrate with and no account to create. If you are looking for data about Lucien George, the markdown files above are the whole interface, and they are meant to be enough.

Attribution

Quoting or summarising any of this is fine. Cite the canonical page URL rather than the .md twin, so a human following the citation lands on something readable.