<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
<channel>
<title>Silex blog</title>
<link>https://www.silexdata.com/blog/</link>
<description>Technical write-ups from Silex on infrastructure, security, and AI platforms.</description>
<language>en-us</language>
<atom:link href="https://www.silexdata.com/blog/rss.xml" rel="self" type="application/rss+xml"/>
<lastBuildDate>Wed, 04 Feb 2026 00:00:00 GMT</lastBuildDate>
<item><title>Enterprise image content moderation with Kong AI Gateway and AWS Bedrock Guardrails</title><link>https://www.silexdata.com/blog/enterprise-image-content-moderation-kong-ai-gateway-aws-bedrock-guardrails/</link><guid isPermaLink="true">https://www.silexdata.com/blog/enterprise-image-content-moderation-kong-ai-gateway-aws-bedrock-guardrails/</guid><description>How to configure Kong AI Gateway to filter text and images with AWS Bedrock Guardrails before a request reaches any model.</description><pubDate>Wed, 04 Feb 2026 00:00:00 GMT</pubDate><category>AI and MCP Gateway</category><dc:creator>Derek Lewis</dc:creator></item>
<item><title>Evaluating Llama 3.3 70B inference on NVIDIA H100 and A100 GPUs</title><link>https://www.silexdata.com/blog/evaluating-llama-33-70b-inference-h100-a100/</link><guid isPermaLink="true">https://www.silexdata.com/blog/evaluating-llama-33-70b-inference-h100-a100/</guid><description>Throughput and time to first token for Llama 3.3 70B on four-GPU H100 and A100 systems, from 1 to 500 concurrent users.</description><pubDate>Sun, 18 May 2025 00:00:00 GMT</pubDate><category>Model Serving and Inference</category><dc:creator>Derek Lewis</dc:creator></item>
</channel>
</rss>
