<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Uci on Aayush Bajaj's Augmenting Infrastructure</title><link>https://abaj.ai/tags/uci/</link><description>Recent content in Uci on Aayush Bajaj's Augmenting Infrastructure</description><generator>Hugo</generator><language>en</language><copyright>© 2026 Aayush Bajaj</copyright><lastBuildDate>Fri, 10 Jul 2026 08:20:25 +1000</lastBuildDate><atom:link href="https://abaj.ai/tags/uci/index.xml" rel="self" type="application/rss+xml"/><item><title>Chess Bot</title><link>https://abaj.ai/wiki/ai/adv-search/chess-bot/</link><pubDate>Thu, 09 Jul 2026 21:02:56 +1000</pubDate><guid>https://abaj.ai/wiki/ai/adv-search/chess-bot/</guid><description>&lt;p>two chess projects live in this codebase, and honesty requires separating them up front. the first is a &lt;strong>from-scratch javascript engine&lt;/strong> (in &lt;code>arcade/references/chess/js/&lt;/code>, built following the classic bluefever software series): 120-square mailbox board, hand-rolled move generator validated by perft, material + piece-square evaluation, alpha-beta with quiescence, a principal-variation hash table, and iterative deepening. the second is &lt;strong>voice chess&lt;/strong> (&lt;code>code/private/chess-bot/&lt;/code>, published as &lt;a
 href="https://github.com/abaj8494/ollama-voice-chess"
 
 
 class="link--external" target="_blank" rel="noreferrer"
 
>github.com/abaj8494/ollama-voice-chess&lt;/a>): there i did &lt;em>not&lt;/em> write the engine — stockfish plays the moves over uci — and the engineering is in the wrapper: skill throttling, an ollama llm that provides spoken commentary without being allowed to hallucinate moves, neural tts, and spaced-repetition opening training. one project teaches you how engines work; the other teaches you what to build &lt;em>around&lt;/em> an engine.&lt;span class="margin-note" data-note="everything measured on this page — perft counts, engine analysis, commentary — is real output from running the code on this machine.">
 &lt;span class="margin-note-indicator">𐃏&lt;/span>
&lt;/span>
&lt;/p></description></item><item><title>Digits</title><link>https://abaj.ai/tags/digits/</link><pubDate>Thu, 09 Jul 2026 21:02:56 +1000</pubDate><guid>https://abaj.ai/tags/digits/</guid><description>&lt;p>The scikit-learn digits dataset: 1,797 tiny 8x8 greyscale images of handwritten digits, and the workhorse of every sklearn tutorial that needs a multiclass problem which loads instantly and fits in L2 cache.&lt;/p>
&lt;h2 id="provenance">Provenance&lt;a href="#provenance" class="post-heading__anchor" aria-hidden="true">#&lt;/a>
&lt;/h2>
&lt;p>The data is the &lt;strong>Optical Recognition of Handwritten Digits&lt;/strong> (optdigits) set from the &lt;a
 href="https://archive.ics.uci.edu/dataset/80/optical&amp;#43;recognition&amp;#43;of&amp;#43;handwritten&amp;#43;digits"
 
 
 class="link--external" target="_blank" rel="noreferrer"
 
>UCI machine learning repository&lt;/a>, created by E. Alpaydin and C. Kaynak at Bogazici University and donated in July 1998. It originates in Kaynak&amp;rsquo;s 1995 MSc thesis on combining multiple classifiers, and the companion paper is Alpaydin and Kaynak, &lt;em>Cascading Classifiers&lt;/em>, Kybernetika 34(4), 1998.&lt;/p></description></item></channel></rss>