<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Vast-Ai on Kuldeep Pisda</title><link>https://kdpisda.in/tag/vast-ai/</link><description>Recent content in Vast-Ai on Kuldeep Pisda</description><generator>Hugo</generator><language>en-us</language><lastBuildDate>Sat, 25 Apr 2026 10:00:00 +0530</lastBuildDate><atom:link href="https://kdpisda.in/tag/vast-ai/index.xml" rel="self" type="application/rss+xml"/><item><title>Renting Strangers' GPUs by the Minute: The vast.ai Experiment</title><link>https://kdpisda.in/serverless-gpu-transcoding-vast-ai/</link><pubDate>Sat, 25 Apr 2026 10:00:00 +0530</pubDate><guid>https://kdpisda.in/serverless-gpu-transcoding-vast-ai/</guid><description>&lt;p&gt;The free-egress move had rescued the NGO&amp;rsquo;s unit economics and, in the same stroke, handed me back a problem I thought I&amp;rsquo;d permanently outsourced: transcoding. Wasabi wasn&amp;rsquo;t real AWS S3, so AWS MediaConvert couldn&amp;rsquo;t read from it, which meant I was back to encoding video myself with plain &lt;code&gt;ffmpeg&lt;/code&gt; — on CPU boxes that could clear maybe 30 to 40 videos a day against an inbound firehose of 150-plus. I needed GPU-speed encoding. I owned zero GPUs. And I worked for an organization that had just spent a week panicking about a four-figure CDN bill, so &amp;ldquo;let me buy a rack of NVIDIA cards&amp;rdquo; was not a sentence I could say out loud in a budget meeting.&lt;/p&gt;</description></item></channel></rss>