Posts on Xianzhi: Portfolio Site

Solving a Saddle Point Problem

Wed, 13 May 2026 11:07:06 -0500

A finite-element saddle-point solver written with deal.II

This post walks through a research/coursework project that I worked on this during spring 2025 semester: a two-component finite-element solver, written in C++ using the deal.II library, for a saddle-point system that arises when one tries to compute the harmonic Riesz representation of a given functional in the dual of $H^1$. The full code documentation lives at saddle-point-project url, it is grown out of the MATH 676 project that I completed in spring 2025.

Tech Notebook: A Machine Learning & Deep Learning Journey

Sat, 03 Jan 2026 22:34:22 -0500

The actual blog (with code) is at: https://xianzhiwang1.github.io/tech-blog/

I maintain a personal technical blog where I write up what I learn as I work through machine learning and deep learning, both inside coursework and on my own. This project grows out of the CS0451 Machine Learning class I took at Middlebury College. The overarching theme of this tech blog is understanding by building: each post takes a single idea, walks through the math behind it, implements it in Python (usually from scratch first, then alongside a library reference), and then puts the implementation to work on a real dataset. The site is built with Quarto and published from a stack of Jupyter notebooks, so every post is a fully reproducible mix of derivations, explanations, code, plots, and embedded source files.

Neovim (Lua) authoring: Packer.nvim, Markdown preview, and TeX Live

Tue, 14 Jan 2025 12:54:45 -0600

Neovim is my primary modal editor for technical writing and lightweight IDE-style tasks. Configuration is expressed in Lua and versioned in this repository. The setup is intentionally incremental: curated plugins, predictable keymaps, and a reproducible local toolchain rather than a monolithic distribution.

Plugin management uses Packer.nvim today; I am evaluating Lazy.nvim for faster startup and declarative spec loading (see below). For Markdown, Alt+k triggers markdown-preview.nvim in an existing browser session—equivalent to :MarkdownPreview—via Packer declarations such as:

Terminal email with Mutt: Gmail, OAuth 2.0, IMAP, and SMTP

Tue, 14 Jan 2025 12:51:24 -0600

This write-up summarizes a CLI-first email stack built around Mutt on Linux, with Google Gmail over IMAP/SMTP and OAuth 2.0 for authentication—relevant when Google Workspace or policy blocks legacy app passwords.

For a personal Gmail account where app passwords remain available, mutt-wizard plus an app password is the fastest path: generate credentials in the Google Account security UI, install the wizard from GitHub, and follow its prompts. That flow assumes comfort with the shell and a working muttrc layout.

Layered keyboard layout with KMonad on Linux (i3)

Tue, 14 Jan 2025 12:48:18 -0600

This note documents how I use KMonad, a cross-platform, user-space keyboard remapping daemon written in Haskell, to implement a layered keymap on a Linux workstation. The configuration lives alongside my window manager setup in this i3 repository.

Compared to QMK/ZMK firmware, KMonad runs in user space on the host OS, which makes it easy to iterate on declarative layout definitions without flashing hardware. The mental model is similar to embedded keymaps: layers, tap-hold behaviors, and modifier chords, expressed as structured configuration rather than ad hoc desktop shortcuts.